Content latent memory analysis method

Abstract

To provide a content latent memory analysis system capable of collecting individual reactions according to a fixed schedule and accurately confirming that a high level of transfer memory has been reached.SOLUTION: A learner terminal 10i, a web-side server Wb, and a local-side server Wc are connected via a communication network, and a presentation list of a learner generated based on schedule conditions is sent to the learner terminal, and the learner's reactions to the content are collected, and latent memory is analyzed based on these reactions.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a content implicit memory analysis method. [Background technology]

[0002] In recent years, it has become popular to learn English vocabulary through e-learning. However, when learning English vocabulary, which requires continuous and repeated study over a long period of time, even with this type of repeated study, students may end up feeling that "I can't remember it no matter how many times I try" or "I wonder if I can really remember it," and eventually give up on studying.

[0003] In other words, if students cannot see the effects and results of their learning, their motivation to learn will certainly decrease. A learning method to solve this problem is disclosed in Patent Document 1 (Patent Publication No. 5130272), for example.

[0004] Patent Document 1 is a method for predicting learning effectiveness that uses a schedule creation method to collect individual responses according to a set schedule, enabling analysis and prediction that takes into account individual attributes and the quality of the content. This is significantly different from traditional drill learning, in which questions are given and students are instructed to "solve and memorize," and it can measure whether knowledge has been acquired at a long-term memory level (something seen once remains in the memory at a certain level for a set period of time), or in other words, whether the knowledge has become "usable."

[0005] The learning effect prediction method of Patent Document 1 defines a schedule for presenting learning content (e.g., English vocabulary) to a learner using schedule conditions that combine timing conditions (what time) and presentation conditions (difficulty level, content type, number of repetitions, etc.).

[0006] Then, a first learning schedule is generated that defines a first interval condition for presenting the content at intervals (e.g., 4 days) shorter than a predetermined period (e.g., 5 days), and a second learning schedule is generated that defines a second interval condition for presenting the content at intervals (e.g., 6 days) longer than the predetermined period (e.g., 5 days).

[0007] Then, a presentation list is generated based on these learning schedules (first and second) for presenting content to multiple learners under first interval conditions and second interval conditions, and the content is presented to the predicted target learner under the first interval conditions based on the learning schedules.

[0008] Then, performance data (poor, not good at all, a little better, good: check the level of confidence in one's understanding) of the multiple learners when they continue learning using the presented content under the first interval condition and the second interval condition is collected.

[0009] Based on this collected performance data, the learner's performance as they continue to study is predicted. The timing conditions described above are based on the technology disclosed in Japanese Patent No. 3764456.

[0010] In other words, in order to make events related to all contents occur repeatedly at a fixed timing, an event cycle unit is set as a fixed period that is longer than the presentation unit, which is the minimum period for event occurrence, and shorter than the interval between the start of an event and the occurrence of the next event, and the contents are arranged so that a specific event related to each content occurs once within that event cycle unit. [Prior art documents] [Patent documents]

[0011] [Patent Document 1] Patent No. 5130272 [Patent Document 1] Patent No. 3764456 Summary of the Invention [Problem to be solved by the invention]

[0012] In Patent Document 1, when a learner sends a personal assessment (also called a grade assessment) that one of the words is "good: means that it is understood," the presentation of that word will be discontinued from that stage onwards in the learner's schedule TB (including when "good" is selected by mistake).

[0013] However, it is relatively easy for humans to remember things that happened the day before or just a few days ago. This memory is called explicit memory (specifically, the memory of what was for dinner the day before yesterday). Therefore, if you study a lot of words at the last minute (cramming) in an attempt to improve your grades, you will rate those words as "good (meaning you understand them)" even if you were unable to do so several weeks or months ago.

[0014] On the other hand, humans remember things they have seen once for several months (even if they have only seen something for a few seconds, the information is still retained in their brains for several months). Although it depends on the material (for example, the difficulty of the English words), if the content is stored at a level of ability (latent memory: knowledge acquired at a level of ability), the content can be retrieved immediately.

[0015] In other words, in subsequent learning, even if the word is not implicitly memorized, the content (problem) will be eliminated.

[0016] Therefore, even if an individual learner has not yet reached a predetermined level of long-term implicit memory [how much learning (e.g., vocabulary) has been acquired], they will not be allowed to study that word again, which creates a serious problem: they will not be able to reach the predetermined level of implicit memory for that word.

[0017] The present invention has been devised to solve the above-mentioned problems. [Means for solving the problem]

[0018] The implicit memory analysis method of the present invention is a method of analyzing implicit memory, which connects a learner's terminal and a server at an analysis center via a communication network, and transmits to the learner's terminal a presentation list and a content list for each learning session, which are generated based on schedule conditions for learning learning content corresponding to the learner's attribute information at regular intervals over a learning period, and analyzes the learner's implicit memory achievement level for the learning content based on the learner's response data to the learning content included in the response list (transmission instruction) from the learner's terminal, It is a method of describing experiences (a set of consecutive events) that precisely describes the expression of "who," "when," "what," and "how" an action was taken, and the expressions of "when" and "how" are expressed as timing conditions based on whether or not a specific event occurred within a presentation unit, and the number of times the event occurred within the presentation unit is separated into a presentation condition called the number of repetitions, thereby simplifying the expression of "when" and at the same time making the expressions of "when" and "how" independent. The presentation unit is a minimum period of the schedule condition, and the presentation list is associated with this presentation unit to simplify the expression of "when," and further, the presentation condition is defined in the presentation list to separate the expressions of "when" and "how," The analysis center transmits to the learner terminal a task process that executes a process that specifies which content elements to present, in what order, and how to record responses for a set of various content elements, and then includes the task code in the presentation conditions and transmits the same to the learner terminal; Display the content on the learner's terminal and acquire the response, and if the highest score is obtained, assign a skip code to skip displaying and recording the subsequent same content in the presentation list; The analysis center receives the response list for the learning session, and if the response is the highest point for a predetermined number of consecutive sessions compared with past response lists and the regression line for this learning session reaches the highest point, assigns a skip code to the same learning content in the presentation list for each learning session over the subsequent learning period; This is transmitted to the learner's terminal as the presentation list for each learning session. Furthermore, the present invention provides a method for analyzing implicit memory, which connects a learner's terminal with a server at an analysis center via a communication network, and transmits to the learner's terminal a presentation list for each learning session and a content list by difficulty level, which are generated based on schedule conditions for learning learning content corresponding to the learner's attribute information at regular intervals over a learning period, and analyzes the learner's implicit memory achievement level for the learning content based on the learner's response data to the learning content contained in the response list from the learner's terminal, The presentation list includes: The presented data, which is a set of a serial code of the learning content in the content list, a column for acquiring an individual evaluation score, a column for acquiring a reaction time, and a task code of the TASK process, is arranged for the same learning content and different learning content, The learner terminal includes: (A1). The TASK process is composed of multiple tasks that are executed without separating how to display the learning content and what data to acquire according to the set task code; (A2) skip task processing for skipping the designated presentation data and designating the next presentation data; (A3) A process of reading the reaction time acquisition column with priority, and if this column is "empty", setting the task code to TASK processing and executing it, and if a negative value is written, setting a skip code to execute the skip task processing; (A4) A process of using a set of the serial code of the presentation data in the presentation list, the individual performance evaluation value in the column for obtaining the individual performance evaluation value, and the reaction time in the column for obtaining the reaction time as reaction data, and transmitting this reaction data as a reaction list to the analysis center in response to a transmission instruction; is downloaded from the analysis center's server, The negative value is When the skip task processing on the learner terminal writes the highest score in the column for obtaining the individual evaluation score, a negative value is written in the column for obtaining the reaction time of all subsequent presentation data having the same serial code in the presentation list, and further, The server of the analysis center A method for analyzing implicit memory, characterized in that each time a response list for each learning session of the learner is received, a response list up to a predetermined number of sessions prior is specified, and if the individual performance evaluation value in the column for obtaining individual performance evaluation value of response data having the same serial code in these response lists shows the highest score consecutively and the regression line of the learning content of this serial code reaches the highest point, it is determined that sufficient implicit memory has been achieved and a negative value is written in the column for obtaining reaction time of presentation data having the same serial code in all subsequent presentation lists, and this is sent to the learner's terminal as a presentation list for each learning session. [Effects of the Invention]

[0019] According to the present invention, responses from individuals are collected according to a set schedule, and it is possible to accurately confirm whether a sufficient level of implicit memory has been achieved. Therefore, it is possible to obtain a schedule condition for each individual in which content that has truly reached a sufficient level of implicit memory is deleted from subsequent learning. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a schematic configuration diagram of a content latent memory analysis system according to an embodiment of the present invention; [Figure 2] This is an explanatory diagram of implicit memory. [Figure 3] FIG. 10 is an explanatory diagram of a content DB. [Figure 4] FIG. 10 is an explanatory diagram of a schedule table. [Figure 5] FIG. 10 is a specific explanatory diagram of a schedule TB based on empirical schedule conditions. [Figure 6] FIG. 10 is an explanatory diagram of a local side presentation list TRi and a web side presentation list WRi. [Figure 7] 1 is an explanatory diagram of the web-side prepared presentation list SRi, the terminal-side generated one-day presentation list SFi, and the personal evaluation response list SHi. [Figure 8] FIG. 10 is an explanatory diagram of a grade definition coordinate system HZi that defines the grade rating Gmi. [Figure 9] FIG. 10 is an explanatory diagram of a coordinate system Vci for confirming latent memory that defines the individual's grade point average Pm. [Figure 10] FIG. 10 is an explanatory diagram of an outline of an application on a learner terminal. [Figure 11] FIG. 10 is an explanatory diagram of the individual evaluation response list SHi. [Figure 12] This is a sequence diagram (1) of the content implicit memory analysis system. [Figure 13] This is a sequence diagram (2) of the content implicit memory analysis system. [Figure 14] FIG. 10 is an explanatory diagram of a presentation file GRi to be sent to a learner. [Figure 15] This is a sequence diagram (3) of the content implicit memory analysis system. [Figure 16] FIG. 1 is an explanatory diagram (1) of the screen of a learner's terminal when logging in. [Figure 17] FIG. 2 is an explanatory diagram (2) of the screen of the learner's terminal when logging in. [Figure 18] 1 is a flowchart (1) illustrating a learner's presentation preparation process. [Figure 19] 10 is a flowchart (2) illustrating the learner's presentation preparation process. [Figure 20] 1 is an explanatory diagram (1) of a specific example of display processing in the learner terminal 10. FIG. [Figure 21] FIG. 10 is an explanatory diagram (2) of a specific example of display processing in the learner terminal 10. [Figure 22] FIG. 1 is an explanatory diagram of a task. [Figure 23] This is a sequence diagram (4) of the content implicit memory analysis system. [Figure 24] 10 is a flowchart of a web-side content performance value definition process. [Figure 25] 10 is an explanatory diagram of the definition of a specific coordinate system Vci for confirming the potential memory level of the individual's grade point average Pm. [Figure 26] FIG. 10 is an explanatory diagram of the association between the response list Hi and TRi of the local server. [Figure 27] FIG. 10 is an explanatory diagram showing a regression line in the form of a table. [Figure 28] This is an explanatory diagram of the generation of a one-day presentation list SFi for terminal side generation by a web-side application server. [Figure 29] This is an explanatory diagram of the processing when the web-side application server receives the response list Hi. [Figure 30] FIG. 10 is an explanatory diagram of the effects of different schedule conditions (timing conditions). [Figure 31] FIG. 1 is an explanatory diagram of a presentation unit, an interval, and an event cycle unit. [Figure 32] FIG. 10 is a diagram showing the homogenization of presentation conditions and condition units and presentation units. [Figure 33] FIG. 1 is an explanatory diagram of a test configuration. [Figure 34] FIG. 10 is a diagram showing an example of a schedule condition including a schedule expressed by a basic unit and a sub-unit. [Figure 35] FIG. 1 is a diagram showing an example of a content table. [Figure 36] FIG. 2 is a diagram showing an example of a content table. [Figure 37] FIG. 1 is an explanatory diagram (1) of a specific example of a schedule table. [Figure 38] FIG. 10 is an explanatory diagram of another specific example of the content table. [Figure 39] FIG. 10 illustrates timing conditions and content item allocation. [Figure 40] FIG. 10 is a diagram illustrating the effect of a regression line. DETAILED DESCRIPTION OF THE INVENTION

[0021] In this embodiment, the content to be studied can be almost any type of learning content, including kanji, English words, idioms, academic terms, medical terms, pharmaceutical terms, etc., but in this embodiment, it will be described as English words (also called words). Furthermore, "study" includes learning through drill questions and learning through tests. The content will also be simply called words.

[0022] Humans remember things they have seen once in their brains for several months (even if they have only seen something for a few seconds, the information is still retained in their brains for several months: implicit memory). The time series changes (fluctuation patterns) corresponding to the learning of this latent memory (also referred to as mastery or familiarity in this embodiment) differ depending on the material (e.g., the difficulty of the English words) as well as the level of mastery.

[0023] For example, as shown in Figure 2, when a word is encountered for the first time (including the answer), the rate immediately drops to about 20%, and then the curve maintains a certain level (La).

[0024] Furthermore, when students look at words that they know a little (have learned) or simple words, the curve (Lb) shows a decline from around the 5th day, reaching about 40% by the 10th day. When students learn even easier words (for example, when they look at the answers and words) the level of implicit memory (true ability) is about 90% for about 10 days, and by the 30th day the curve (Le) shows about 40%. This embodiment is a method for describing experiences (a set of consecutive events) by precisely describing "who," "when," "what," and "how" an action was taken. The above expressions of "when" and "how" are expressed as timing conditions based on whether or not a specific event has occurred within a presentation unit, and the number of times the event has occurred within the presentation unit is separated into a presentation condition called the number of repetitions. This method simplifies the expression of "when" and at the same time, a program is used that separates the expressions of "when" and "how." (ii) It is also a way of describing experiences (a set of consecutive events) by precisely describing the "who," "when," "what," and "how" of an action. As a method for separating the expressions of "how" and "what," a task process (execution program: TASK "0") that specifies which content elements to present, in what order, and how to record responses for a collection of various content elements such as text, audio, video, and images, and executes the process, and the task code is used to: A method (program) is used that strictly specifies "what" and "how" an action was taken.

[0025] (iii)(i) and (ii) correspond to the principle that allows the same processing to be achieved even when switching content databases. The content code and execution program separate the collection of content elements (English words, Chinese, French, Japanese, etc.) from "how" they are to be presented and recorded, thereby separating the collection of various content elements (called content) from the presentation and recording process (called task processing), and is a method (program) for executing presentation conditions that allows a series of similar processes to be specified and executed for content even if the content is replaced. For example, if you have content (a collection of content elements) whose elements are English word text and its Japanese translation, and content whose elements are Chinese text and its Japanese translation, you can prepare data with a structure that maps the English word text and Chinese text to a variable called QUESTION, has columns (for example, a QUESTION column and an ANSWER column) that map the Japanese translation text to a variable called ANSWER, and has a column for recording responses (RESPONSE column), and even if the content database (content elements such as English words, Chinese, French, or Japanese) changes, you can express "what" and "how" an action was taken through the content and task processing, and collect responses.

[0026] This allows various presentation conditions to be defined even when switching between various content databases. RESPONSE may also include a time column. A specific example will be described below.

[0027] FIG. 1 is a schematic diagram of a content implicit memory analysis system according to this embodiment. As shown in FIG. 1, the content implicit memory analysis system will be described as using a learner's tablet, smartphone, personal computer, or the like (hereinafter collectively referred to as learner terminal 10i).

[0028] This learner terminal 10i (Android 1, iPhone, etc.) is connected to the service center side (hereinafter referred to as the center side (Wa)) via a communication network 50 such as the Internet network or a VPN network, as shown in FIG.

[0029] The center side (Wa) is made up of a web side (Wb) and a local side (Wc), etc. The local side (Wc) is connected to the web side (Wb) by a LAN 80 (local network).

[0030] The web side (wb) is made up of a web server 100, an application server 200, a web-side database server (hereinafter referred to as a web-side DB server 300), etc. (collectively referred to as a web-side server system).

[0031] On the other hand, the local side (Wc) is made up of a local server 400, a local DB server 500, and the like. These servers are collectively referred to as an analysis center or an analysis center server system. Here, an outline of the content implicit memory analysis system of this embodiment will be explained. For example, the "what" in "what" is the unit of memory, not just memory. The unit of memory is content (also called learning content).

[0032] For example, if there is content (a collection of content elements) whose elements are English word text and Japanese translation text, and content whose elements are Chinese text and Japanese translation text, as mentioned above, by preparing data with a structure that maps the English word text and Chinese text to a variable called QUESTION, has columns (for example, a QUESTION column and an ANSWER column) that map the Japanese translation text to a variable called ANSWER, and has a column that records responses (RESPONSE column), then even if the content database changes, the content and task processing can be used to express "what" and "how" an action was taken, and responses can be collected.

[0033] This allows you to define different presentation conditions even when switching between different content databases. The individual content changes depending on the combination of "what information" and "how conditions." Therefore, the "what" and "how" parts are assumed to be inseparable. TASK processing (a program that specifies TASK processing) specifies how in detail the content "memory" is displayed and how the data is acquired (in what order it is displayed, and at what timing what kind of response is acquired), and TASK code is the symbol that represents this.

[0034] The TASK process (the program that defines the TASK process) is also simply called TASK "0" or an existing TASK or a display control program or a display task. In this embodiment, the presentation conditions are defined in a presentation unit that is separate from the timing of the schedule conditions that trigger learning at a certain timing, and these presentation conditions are defined in a presentation list. The presentation list contains multiple presentation data sets that include a serial code of the learning content, a column for acquiring the individual's evaluation score (response column), a column for acquiring reaction time (time column), and a task code (D, T, etc.) for TASK processing.

[0035] By entering this TASK code and the serial number of the content into the presentation list, the learner's terminal determines how to display the content and the flow of data acquisition. The program (TASK "0") that specifies the TASK processing is downloaded from the learning server (analysis center) when a specified URL is opened in the browser on the learner's device. After that, when a specified TASK processing is requested (selected) by the TASK code on the learner's device, the program (task) is launched and the content display and data acquisition process is carried out.

[0036] In other words, TASK "0" is a process (script) consisting of multiple types of tasks that execute without separating how to display the learning content and how to acquire the data according to the set task code. The learning content display task (TASK "0") specifies the presentation data that constitutes the presentation list for the received learning session (1st, 2nd, ...: 1st day, 2nd day, ...), displays the learning content and time bar, etc., prompts the user to enter their own performance rating value and writes it in the response box (also called a column), and also writes the measurement time in the time box (also called a column) to specify the next presentation data.

[0037] Then, by selecting the send button, these are sent as a reaction list to the analysis center by the sending process. Note that the learning content display task (TASK "0") has processes such as D and T, which will be described later, and the display and collection methods differ. If the personal grade value written in the column for obtaining the personal grade value (response column) for the learning content displayed according to TASK "0" is the highest score (e.g., 3), the terminal-side skip task (TASK "1": (script)) writes a negative value (e.g., "-1") in the time box (column) of all subsequent presentation data having the serial code of the same learning content in the presentation list for the received learning session, and thereafter prioritizes reading the time box (column) and skips displaying this same learning content.

[0038] On the other hand, the analysis center performs analysis by executing a center-side skip task (task "-2" or TASK "2"), which is a function of the individual optimization process. The local server generates and stores a presentation list based on the schedule conditions of the learner. The center-side skip task may be shared between the application server and the web server, or may be performed by the local server wc. In this embodiment, an example in which the center-side skip task is performed by the application server and the web server will be mainly described.

[0039] The center-side skip task of the application server causes the web server to send the login information from the web server to the learner's terminal, and each time the application server receives a response list to this presentation list, it determines whether the learning content has shown the highest score three times in a row. The application server also executes the center-side skip task. This center-side skip task calculates the fluctuation in grades for each learning content, applies a regression line, for example, as a predictive function, and calculates whether the value of this regression line exceeds the intercept, slope, and maximum score (3). This information, along with the judgment result that the maximum score has been exceeded, for example, three times in a row, is recorded in a CSV file or the like along with the serial code and uploaded to a web server.

[0040] The web server finds the learning content that is estimated to have exceeded the highest score from the application server's calculation results (CSV) in the presentation list (multiple learning sessions) for the learning period (for example, 60 weeks) prepared on the local server, and records (-2) in the time column of each serial numbered problem. In subsequent learning sessions, by preparing a program (script) that does not display problems with a (-2) recorded on the device, it is possible to create a situation where problems that have been cleared are not displayed to the learner.

[0041] In other words, each time a reaction list for each learning session of the learner is received, the reaction list from the specified previous session (three sessions ago, including the current one) is specified, and if the individual score evaluation value in the column for obtaining the individual score evaluation value of reaction data with the same serial code shows the highest score (the highest score is used as an example) consecutively, and the regression line of the learning content of this serial code reaches the highest point, it is determined that sufficient latent memory has been reached, and a negative value (-2) is written in the reaction time acquisition column of presentation data with the same serial code in all subsequent presentation lists (center-side skip processing), and this is sent to the learner's terminal as a presentation list for each learning session.

[0042] (Web side (wb)) The web server 100 has the function of transmitting the learning terminal side application (including the learning content display task (TASK "0"), the terminal side skip task, etc.: described later) stored in the web side DB server 300 to the learner terminal 10i (10a, 10b, ...) for downloading, and then providing display information such as HTML and images in response to a web browser request from the learner terminal 10i.

[0043] In this embodiment, after the learner terminal side application described later is provided from the application server 200 to the learner terminal 10i, a presentation list of the learning session and learning content (for example, one day's worth, five days' worth, or one month's worth) are provided (transmitted), and the reaction data hi of the reaction list from the learner terminal 10i is stored in the web-side DB server 300.

[0044] Furthermore, a presentation file GRi to be sent to the learner, which will be described later and which is generated by the application server 200, is sent in HTML to the learner terminal 10i. Before explaining the application server 200 and the web-side DB server 300, the local side will be explained.

[0045] (Local side (wc)) The local side (wc) is made up of a local server 400, a local side DB server 500, etc. (collectively referred to as a local side server system). The local-side DB server 500 includes a schedule TB memory 510 storing a schedule table (hereinafter referred to as a schedule TB (e.g., Excel format)), a content DB memory 520 storing a content database (hereinafter referred to as a content DB: CSV format), a presentation list memory 530 storing a presentation list (local-side presentation list TRi: CSV format) etc. It also includes a memory (not shown) for storing learner information Ai etc.

[0046] This content DB is composed of, for example, a collection of content elements of English word text and a collection of content elements of Japanese translation text, with English words associated with Japanese and Japanese words associated with English words. In addition, the content DB can be exchanged (switched) according to Chinese, French, German, medical terminology, national medical examination terminology, four-character idioms, etc.

[0047] The learner information Ai (A1, A2, ...) includes the name of the contracting school (which may be an individual), age, school name (also called group name), gender, learning course (e.g., English or Japanese for first-year junior high school students), learning level (beginner, intermediate, advanced), contract details (learning schedule, fee), name of the instructor, learner terminal address, email address, telephone number, date and time of registration, etc. To prevent leakage of personal information, it is common practice not to enter personally identifiable information into the system or to delete such information on the Internet (in this embodiment, notations such as name and school name are used for clarity). The registration of this learner information Ai will be described later.

[0048] The learner information Ai (A1, A2, . . . ) is stored by group (for example, by school). The learning level and learning course correspond to the difficulty level N of the content. The learner terminal address, email address, and telephone number are collectively referred to as learner terminal information Di (D1, D2, . . . ).

[0049] The local DB server 500 stores the ID code, password (account), registration number, and registration date and time by year in association with each other. The ID code, password (account), group name, learner information registration number, difficulty level N, etc. are collectively referred to as learner basic identification information Bi (B1, B2, ...: account, group name, name, registration number, difficulty level N: unique code, preferably).

[0050] In addition, the learning terminal information Di (learner terminal address, email address, telephone number) and the learner basic identification information Bi are collectively referred to as learner-specific identification information SGi (SG1, SG2, ...: name, group name, difficulty level N, account, registration number, learner terminal address, learner terminal number, email address, telephone number: unique code is preferred).

[0051] The learning course (for example, English or Japanese for first-year junior high school students) and learning level (beginner, intermediate, advanced) correspond to the difficulty level N. A learning schedule is, for example, one lesson (or multiple lessons) per day (or two, three, or four days, etc.) for, for example, 60 weeks (analysis period), with a short test once a week, for example. It is preferable that the learning course and learning level be decided in consultation between the teacher in charge and the learner.

[0052] [Content DB] Fig. 3 is an explanatory diagram of a content DB. As shown in Fig. 3(a), the content DB stores content data in order of difficulty level N from lowest to highest. This collection of content data CDi (CD1, CD2, ...) is called the content DB.

[0053] Specifically, as shown in Figure 3(a), the content DB consists of a content number ("001": a serial code), content [vocabulary (questions and answers)], difficulty level N (for example, level 1 beginner, intermediate, advanced for first-year junior high school students, level 1 beginner for second-year junior high school students, etc.), and content DB type, and is further associated with a schedule condition SJ (identification code). The content DB type indicates what type of learning (D, T, etc.) the content is used for.

[0054] This schedule condition SJ is composed of learner classification information indicating what kind of course it is for and what level of people it is for, content type (content set), timing conditions, presentation conditions (including display task code and number of repetitions), etc.

[0055] The display task code (also simply referred to as a task code or a TASK "0" code) may be, for example, D (drill) or T (test), . . . Other than these, there are K (e.g., drills for first-year high school students), U (e.g., tests for first-year high school students), O (questionnaires), X (instructions), F (familiarity assessments), R (recognition tests), etc.

[0056] The timing conditions for these schedule conditions SJ are to study once on the first day (1-day interval), once every 2 days (2-day interval), once every 3 days (3-day interval), and once every 10 days (10-day interval).

[0057] The number of times the presentation condition is repeated is, for example, the same content once per day (number of times the content is presented), the same content once every two days, ..., the same content three times per ten days (number of times it is displayed on the screen).

[0058] The aforementioned display task (simply a task, or a display control program, or TASK "0") is a display control program (also called display control data) that defines (specifies) how to display content and obtain responses according to a task code (TASK "0" code), and is identified (also called a display control code) by a task code (e.g., D, T, ...).

[0059] This display task (display control program or TASK "0") generally displays learning content, prompts the user to input their own grade (0, 1, 2, or 3) and write it in a box for obtaining their own grade (response box: also simply called "response"), and also prompts the user to write the measured time (for example, 1.555: serial value) until the user's grade is input in a box for obtaining reaction time (time box: also simply called "time"). This will be described in more detail later. Although the term "box" refers to a column, in this embodiment, the box will be used for explanation.

[0060] In this embodiment, the processing of this display task (simply task or display control program or TASK "0"), TASK "1", and terminal side TASK "2" is stored in the application server 200, and is downloaded from the learning server when a specified URL is opened in the terminal browser.

[0061] 4(a), the schedule conditions SJ are schedule conditions SA, SB, ..., where schedule condition SA means, for example, learning once a day, SB means once every four days, SC means once every two days, and SD means once every five days. Note that learning once means sequentially displaying each content item in a presentation list (for example, 100-200-300 items of content) for one day (also called once: learning session) described below on the screen, and collecting responses to the items.

[0062] Figure 3(b) shows a specific example of a content DB (also called content TB) (more details are shown in Figure 36). This content DB is associated with a code MD indicating "when." It also shows that a schedule condition is associated with it. The schedule condition consists of the difficulty level of the problem, the task code, the number of repetitions, the schedule name, the timing condition, etc. These are specifically the presentation conditions. It is preferable that these content data CDi be updated after analyzing the reaction data hi.

[0063] Based on instructions from the presentation list generating unit 420, the local server 400 transmits (copies) these contents to the web side.

[0064] [Schedule TB] The schedule TB will be described in detail later with reference to the drawings, but before describing the schedule TB, a supplementary explanation will be given regarding the schedule conditions SJ.

[0065] <Schedule Conditions SJ> As shown in Figure 4(a), the schedule conditions SJ are defined by the timing conditions (what time: different for each schedule name), the number of times the content (word) is repeated, the difficulty level N, the schedule name, the task code (meaning the TASK "0" code), and a code (not shown) indicating how many contents can be allocated to one study (first day, second day, ...: study session) (determined empirically: see Figure 4(a)). In Figure 4(a), the TASK "0" code is simply written as "task."

[0066] The schedule conditions SJ are also associated with learner classification information (not shown) indicating what course and what level the course is for, an analysis period (also called a study period), the total number of contents within the analysis period, etc. The schedule conditions SJ will be described in detail later.

[0067] FIG. 4(a) shows schedule conditions SB, SC, and SQ for schedule names B, C, and Q. Based on such schedule conditions SJ, schedule conditions (hereinafter referred to as experienced schedule conditions) are generated for each learner.

[0068] The experience schedule conditions (including the analysis period and the total number of contents) are specifically defined in the contents and learner identification information SGi, as shown in Figure 4(b). The contents are specifically content codes (or serial codes). As shown in Figure 4(b), the experiential schedule conditions consist of learner identification information SGi (learner information identification number: written as individual SG), content (content serial code), timing conditions, presentation conditions, schedule name, etc.

[0069] The presentation conditions include the difficulty level, the number of repetitions, the task code, etc. In Figure 4(b), the TASK "0" code is simply written as "task." This task code is a code indicating a display control program such as K, D, T, U, R, or F (called a TASK "0" code). In addition to this, there is a TASK "1" code (also called a TASK "1" code) and a terminal-side TASK "2" code (also called a terminal-side TASK "2" code).

[0070] FIG. 5 is a specific explanatory diagram of a schedule table TB (also called a schedule TB) based on empirical schedule conditions (more details are shown in FIG. 36).

[0071] As shown in Figure 5, the schedule TB consists of "M01D01", "M01D02", ... (where "M" indicates the month, "D" indicates the day, and "01" means, for example, the first item on a certain date in a certain month: i.e., it corresponds to the timing condition), the schedule name (B or C, A or D, ...), and the number of repetitions (R01 or R02, ...R05, ...: R indicates repetition: the same word is limited to five times).

[0072] However, in FIG. 5, B, C, R01, and R02 are all described as the TASK "0" code, but the TASK "0" code is K, D, . . .

[0073] Furthermore, the schedule TB is made up of records (schedule data SDi) each consisting of a column set including difficulty level N, content (actually an identification number), serial, response, time, etc. The order can also be time, serail (code), response.

[0074] In this embodiment, a set of serial, response, time, and task code (TASK "0" code) is referred to as response evaluation collection data HGi. The set (set of records) of the schedule data SDi (SD1, SD2, SD3, . . . ) described above is called a schedule table TB.

[0075] However, since (-1) is written to time on the learner terminal side, in this embodiment, (-1) is also referred to as the learning terminal side determined display skip coefficient αi (also called the first abnormal value or the first negative value).

[0076] In addition, since (-2) is written to time by the web-side skip display control program (center-side skip task), it is also called the web-side determined display skip coefficient βi (also called the second abnormal value or the second negative value). The experience schedule conditions are determined so that if the analysis period is, for example, 60 weeks (which may differ depending on the course), all the contents are allocated in 120 days. In addition, (-1) and (-2) are collectively referred to as abnormal values ​​or negative values.

[0077] For example, a schedule in which students study for four days (equivalent to one week) and take a one-off test (also called a confirmation test) on the fifth day is called one stage. One month or one day can also be considered one stage.

[0078] In addition, the experiential schedule conditions may be generated based on schedule conditions (a set of consecutive events) that mix schedules, for example, to study a certain content (event) on the second day or to study a certain content on the fourth day.

[0079] [Local presentation list TRi] The local presentation list TRi is shown in Fig. 6(a), which is generated based on the experience schedule conditions. Records such as "MD", the word (desk), the experience schedule condition (identification code), and the display control data HGi for collecting reaction ratings (including the course name: not shown) are referred to as local side presentation data Tri.

[0080] A collection of the local-side presentation data Tri (TR1, TR2, ...) is a local-side presentation list TRi for each learning session. In other words, the local-side presentation list TRi (see FIG. 6(a)) is generated for each course, for each learner, and for each learning session. Fig. 6(b) shows the web-side presentation list WRi [WR1, WR2, ...: serial, word (no answer), response, time, task (task code), SGi] generated on the web side. The generation of the web-side presentation list WRi will be explained in the sequence diagram described later.

[0081] [Local Server] As shown in FIG. 1, the local server 400 includes at least a schedule table generating unit 410, a presentation list generating unit 420, a content extracting unit 430, and the like. The schedule table generation unit 410 generates a schedule TB based on experiential schedule conditions according to the difficulty level N, such as a learning course (e.g., English or Japanese for first-year junior high school students) or a learning level (beginner, intermediate, advanced), and stores the schedule TB in the schedule TB memory 510 of the local DB server 500. For details of the schedule TB, see FIG. 5.

[0082] The presentation list generation unit 420 sequentially designates the learner information Ai, generates a local side presentation list TRi (TR1, TR2, ...: see Figure 6(a)) for each learner over the analysis period (e.g., 60 weeks), and stores it in the presentation list memory 530 of the local side DB server 500.

[0083] The presentation list generation unit 420 has a local side presentation list generation process, and sequentially reads the schedule data SDi (SD1, SD2, ...) of the schedule TB in the schedule TB memory 510, and each time this schedule data SDi is read, it reads one content (word) corresponding to the difficulty level N, etc. contained therein, for example, randomly from the content DB in the content DB memory 520 (for example, desk (number 111)).

[0084] Then, "MD" and the word (including the answer: desk: specifically, the word code) contained in the schedule data SDi are read. At this time, a serial code is generated using a random number generator or the like, and the serial code is associated with the schedule data SDi together with a code indicating the time box (also simply referred to as time) and the response box (also simply referred to as response). It is also preferable to associate the serial code with the learner identification information SGi (learner identification information Bji, learning terminal information Di, etc.).

[0085] The time box (also called the reaction time acquisition box) and the response box (also called the individual evaluation score acquisition box) are collectively called history boxes (also called history task boxes) or reaction boxes. In other words, a history box (history task box) refers to a box in which a task (task code) assigned in response to an analysis of a learner's reaction history is written, and in the following, the history box (history task box) will be described as a box for obtaining reaction time and a box for obtaining individual evaluation score values.

[0086] The serial (serial code: word number) may be generated using the learner identification information SGi of the schedule data SDi, the date and time, the identification code of the schedule condition SJ, etc. The learner identification information SGi will also be associated.

[0087] The "MD" in Figure 6(a), the word (desk), the schedule condition SJ (including difficulty level), the serial code, the time box, the response box, and records such as tasks (including course names: not shown) are referred to as the local side presentation data Tri.

[0088] A collection of the local-side presentation data Tri is a local-side presentation list TRi. That is, the local-side presentation list TRi is generated in the presentation list memory 530 for each course and each learner.

[0089] Then, the web side presentation list sending unit (not shown) of the local server 400 generates a web side presentation list WRi [WR1, WR2, ...: Serial, word (no answer), response, time, task: Figure 6(b)] for each learner identification information SGi and sends it to the web side (Wb).

[0090] Furthermore, the presentation list generating unit 420 performs a local web side presentation list generating process. The local Web-side presentation list generation process allocates (specifies) the local-side presentation list TRi of the learner-by-learner identification information SGi.

[0091] Then, the local web presentation list generation process extracts information including the serial (serial code), time box (specifically, a code indicating a box where time can be written), word (for example, desk (no answer: word code 111)), response box (specifically, a code indicating a box where results can be written), MD (date: timing), task, etc., contained in the allocated local presentation list TRi as web presentation data Wri, and creates a set of web presentation data Wri (for 60 weeks) as a file called web presentation list WRi, stores it in memory, and sends it to the web side (Wb) (see Figure 6(b)).

[0092] That is, on the local side, a web presentation list WRi is generated for each learner identification information SGi for each course and transmitted (copied) to the web side (wb). Furthermore, the content retrieval unit 430 of the local server 400 retrieves content from the local DB server 400 .

[0093] (Content extraction process) The content extraction process searches the content DB for content included in the local presentation list TRi [(TR1, TR2, ...: by course, difficulty level, by learner identification information SGi: 60 weeks' worth) stored in the local DB server 500, and extracts (copies) this searched content as actual content Cmi [(for example, English word (desk) and answer (desk), English word (Apple) and answer (apple), ...)]. Since the same content is not extracted, the total is 120 days' worth.

[0094] These contents are called actual contents Cmi (serial, content number, English word, answer, schedule condition, course, . . . ), and are sent as a file to the web-side DB server 300 .

[0095] [Web DB Server 300] Every time the web DB server 300 receives a web presentation list WRi (see FIG. 6(b)) from the local side (Wc), it stores it in memory and performs processing to generate a presentation list to be sent to the learner.

[0096] (Processing for generating presentation list to be sent to learners) In the process of generating a presentation list to be sent to a learner, a web-side preparation presentation list SRi (see FIG. 7(a)) is generated in a memory from the local side based on the web-side presentation list WRi.

[0097] This web-side preparation and presentation list SRi is shown in Figure 7(a). The web-side preparation and presentation list SRi is a group of records, and is composed of a serial (e.g., 111), a time box, a response box (specifically, a code indicating a box in which the results can be written) (however, depending on the task, some numerical values ​​may be written, in which case the box is set to unwriteable), first web-side preparation and presentation data Srai consisting of MD (year, month, day), etc., and third web-side preparation and presentation data Srci in which (-2) (also called the web-side decision display skip coefficient βi) is written in the time box.

[0098] The second web-side prepared presentation data Srbi will be described later. The third web-side prepared presentation data Srci is generated by a web-side presentation list generation unit 220 of the application server, which will be described later.

[0099] (Web side presentation list generation unit 220 of application server 200) The web-side presentation list generation unit 220 of the application server 200 has a learner presentation list transmission process (not shown). This learner presentation list transmission process reads all of the web-side preparation presentation data Sri (including the first web-side preparation presentation data Srai and the third web-side preparation presentation data Srci) for one day (identified by MD) from the web-side preparation presentation list SRi (for 60 weeks).

[0100] This set of web-side prepared presentation data Sri [Srai, Srci] for one day is called a one-day presentation list SFi generated for terminal side (see FIG. 7(b)).

[0101] Although the serial number is shown as three digits, it may be eight, twelve, or even include English letters.

[0102] That is, a predetermined period into which the analysis period (60 weeks) is divided is considered to be one day, and the list of this one day is the terminal-side generated one-day presentation list SFi. Of course, SFi can also be two days, four days, or one month.

[0103] For the sake of distinction, the web-side prepared presentation data Sri of the terminal-side generated one-day presentation list SFi is called terminal-side generated one-day presentation data Sfi.

[0104] That is, the terminal-side generated one-day presentation data Sfi has web-side pre-latent memory assessment presentation data Sfai (see Figure 7(b)) which includes serial (e.g., "111"), a pose box, a time box (□), and a task code (TASK←TASKE (0:D or K, T, ···)).

[0105] It also has web-side latent memory judged data Sfci (see Figure 7(b)) consisting of serial (e.g., "111"), a pose box, a time box (-2: web-side decision display skip coefficient βi), a task code, etc. Furthermore, the web-side presentation list generation unit 220 reads out the actual content Cmi (e.g., desk) having serial included in the first web-side preparation presentation data Srai and the second web-side preparation presentation data Srci of the terminal-side generated one-day presentation list SFi (FIG. 7(b)) from the web-side content DB. In other words, it reads out one day's worth of actual content Cmi.

[0106] Then, this one day's worth of actual content Cmi is put into a file (hereinafter referred to as one day's content file CFi or simply as content list: content list for the learning session), and combined with a one day's presentation list SFi generated for the terminal side (referred to as one day's learner delivery file GFRi), and sent to the learner's terminal 10.

[0107] Furthermore, the task code of the web-side preparation presentation data Sri of the one-day presentation list SFi generated for the terminal side is associated with the code TASK(0:D or K, T, ...). Also, the code TASK(2) is associated with the task code.

[0108] Here, the display control program (task: TASK "0") will be described. The display control program is associated with a task code such as D, T, or K, etc., and is included in a one-day learner delivery file GFRi and sent (SFi) from the application server 200 to the learner terminal 10i via the web server 240.

[0109] The display control program (TASK "0") sequentially specifies the one-day terminal-side generated presentation data Sfi in the one-day terminal-side generated presentation list SFi of the learner terminal 10i described later, and when it is notified that each specification does not contain any abnormal values ​​(for example, -), it displays the content of this one-day terminal-side generated presentation data Sfi on the learner's screen, writes the performance rating Gmi in the response box (box for obtaining the individual rating score value) of the data Shi for obtaining the individual rating score value (also called the learner-specific content reaction collection data or simply the reaction data) in the individual rating reaction list SHi (see Figure 7(c) and Figure 11) described later for this content, and writes the reaction time in the time box (box for obtaining the reaction time), and when the send button is selected, sends this collection to the center as a reaction list Hi.

[0110] The task definition processing unit of the web side presentation list generating unit 220 includes the above-mentioned task code (D, K, . . . ) or TASK "2" code in the web side preparation presentation data Sri.

[0111] Furthermore, if the latent memory confirmation unit 210 confirms that the learner has a high level of latent memory for the serial code content, the task definition processing unit of the web-side presentation list generation unit 220 has the same serial code content in each of the daily web-side preparation presentation lists SRi (including web-side preparation presentation lists having serial codes that are deemed to have a high level of latent memory) that make up the web-side preparation presentation list SRi (60 weeks: see Figure 7(a)). A second abnormal value (-2) is written in the time box of all the web-side prepared presentation data Sri having the serial code in all the subsequent presentation lists.

[0112] As a result, in the learner terminal 10i, the task code of the terminal-side generated one-day presentation data Sfi is set to time (TASK←time), and the content of (-2) is skipped and removed from the display (not displayed).

[0113] That is, the learner terminal 10i executes the process of terminal TASK "2" (see FIG. 22).

[0114] (Latent memory level confirmation unit 210) The potential memory checking unit 210 performs a web-side content performance value definition process and a potential memory checking process.

[0115] The web-side content grade value definition process specifies the response data hi of the response list Hi from the learner terminal 10, and sequentially defines the serial (content) grade evaluation value Gmi (for example, (0), (1), (2), or (3): also simply called grade) contained therein in the grade definition coordinate system HZi (specifically, a function) (see Figure 8). However, if a negative value (-1 or -2) is written in the time field of the reaction data, this reaction data is skipped and the next reaction data is specified.

[0116] The concept of this grade definition coordinate system HZi will be explained using Figure 8. The grade definition coordinate system HZi is associated with serial and learner identification information SGi. In other words, it is generated for each learner identification information SGi and for each serial (content). Note that the number of times the same content is repeated in one day is limited to five or less, so in this embodiment, four times is shown as an example. Note that it is generated using Excel, which creates records of sets of learning sessions, serial, grade value Gmi (also called individual grade value), learner identification information SGi (identification code), etc., but for ease of understanding, this embodiment will explain it as the grade definition coordinate system HZi.

[0117] As shown in FIG. 8, the performance definition coordinate system HZi has the vertical axis Yi (performance rating Gmi) and the horizontal axis Xai (for example, the first, second, etc. on the first day: learning times). As shown in Fig. 9, the latent memory confirmation process defines the individual's own average grade Pm of the serial code (content) of the specified response data hi in a coordinate system Vci (specifically, a function) for confirming latent memory, with the vertical axis representing the individual's own average grade Pm and the horizontal axis representing the number of presentation cycles Xi (also referred to as the number of presentation cycles, number of events, or number of learning sessions: for example, the first day, the second day, ...). This coordinate system Vci for confirming latent memory is generated for each learner's identification information SGi and for each serial (content).

[0118] The latent memory confirmation process averages the performance rating Gmi defined in the performance definition coordinate system HZi for each presentation cycle number Xi (number of presentation cycles or number of events), and sequentially defines this average value as the individual performance mean rating Pm (e.g., day 1, day 2, etc.).

[0119] FIG. 9 shows that the personal grade average Pm of serial (e.g., 225:temple) is defined as "0: completely bad" on the first day, "3: good" on the tenth day, "2: a little better" on the eleventh day, and all "3: good" on the twelfth, thirteenth, and fourteenth days. Note that the horizontal line of the personal grade average Pm(3) is referred to as the highest dotted line MFi in this embodiment. The personal grade average Pm may be 2.9 to 3.0.

[0120] In the latent memory confirmation process, each time a personal grade average Pm is defined in this latent memory confirmation coordinate system Vci, the learning session (also referred to as the number of presentation cycles Xi) in which this personal grade average Pm was defined is specified, and it is determined whether or not a second condition is met, which is that the highest score ma was "3: Good" (also referred to as the highest grade) for three consecutive sessions (including the specified learning session) prior to the specified number of presentation cycles Xi. If the second condition is met, a grade feature function indicating the distribution tendency of the personal grade average (Pm) is calculated. This grade feature function is a regression line or curve. In this embodiment, a regression line is used.

[0121] That is, the regression line yi (yi=axi+b) is found. After the regression line is found, it may be determined whether the highest point ma is "3: Good" consecutively. Then, it is determined whether the value (yi) of this regression line yi (yi=axi+b) meets the first condition (yi=3), which states that the value (yi) reaches the highest dotted line MFi (yi=3: or higher is acceptable).

[0122] If both the first and second conditions are met (also called the third condition), it is determined that the learner has implicitly mastered the content of the serial (e.g., 225:temple) (has a sufficiently high implicit memory). Then, sfi in the terminal-side generated one-day presentation list SFi that has the serial of the specified response data hi is retrieved, and the second negative value ((-2)::web-side determined display skip coefficient βi) is written in the time box included in it.

[0123] Figure 7(b) shows an example where the serial number is "020", and (-2) is written in the time box (□) of this "020" (sfci). Note that (-2) is just an example, and it could also be -3, -4 or -5.

[0124] Furthermore, if it is determined that the first condition and / or the second condition are not met, it is determined that the learner does not yet have a high level of implicit memory for the serial content (e.g., 225:temple). If it is determined that the content pair does not yet have a high level of latent memory, nothing is written to the time of sfi in the terminal-side generated one-day presentation list SFi having the serial of the specified reaction data hi (it becomes sfai).

[0125] Furthermore, each time the individual's average grade Pm is defined in the coordinate system Vci for checking the degree of memory, the regression line yi (yi = axi + b) is calculated, and when this regression line reaches the highest dotted line MFi (highest grade (3)), if the highest score ma is "3: Good" for the previous three consecutive times (including the specified learning session), it can be determined that the learner has implicitly mastered the serial (e.g., 225: temple) content (satisfied the degree of implicit memory or become familiar with it).

[0126] Therefore, even if a learner mistakenly selects "good" a few times before, or if they study overnight, content such as serial (e.g., 225:temple) will no longer be removed from the display, as was the case in the past.

[0127] Then, if it is determined that the learner has implicitly memorized the content of serial (e.g., 225:temple) in his / her brain (it is confirmed that implicit memory has been satisfied), it specifies all subsequent one-day presentation lists SFi for terminal side that have serial (e.g., temple) in the above one-day presentation list SFi for terminal side (see Figure 7(b)), and writes (-2) into the time of the one-day presentation data Sfi for terminal side that have serial (e.g., temple) included in these lists. This process is performed for each reaction data hi in the reaction list Hi.

[0128] Then, when the web-side reaction list transmission process has performed the above process for all the reaction data hi, the reaction list Hi is transmitted to the local server and this reaction list is associated with TRi (see FIG. 26).

[0129] (Student's device) The learner terminal side application of the learner terminal 10i will be further outlined with reference to Fig. 10. As shown in Fig. 10, the learner terminal 10i has a display unit 11, a browser 17, a TASK "0" processing unit 12 (also referred to as a display processing unit), a time task code determination unit 13 (a write data error checking unit), a presentation data memory 14, a presentation data reading unit 15, a terminal side skip unit 16, etc. The browser 17, the TASK "0" processing unit 12, the time task code determining unit 13, the presentation data reading unit 15, and the terminal side skip unit 16 are programs.

[0130] The TASK "0" processing unit 12 is an existing display control program, and includes a learner's side presentation preparation process. This learner-side presentation preparation process copies the form (table, items, learner identification code, etc.) of the individual assessment response list SHi (see Figure 7(c)) which is similar to the terminal-side generated one-day presentation list SFi (Figure 7(b)) contained in the one-day learner delivery file GFRi (one-day terminal-side generated one-day presentation list SFi (Figure 7(b) and one-day learner-specific content file CFi) downloaded from the web into memory. Record numbers are assigned to this table. For example, the task codes "D", "T", etc., and the serial "114" (for example, the word "Temple") are also copied. Furthermore, the learning content is displayed on display unit 11 by browser 17 in a presentation (display) method that corresponds to the task code.

[0131] For example, display English (Temple), and by selecting the answer display button, the answer (Temple) and your own evaluation item (not good, not good, a little better, good) are displayed. By selecting the grade evaluation item Fm, you can write the result (also called a response), and by selecting the send button, you can send it (to the center). The response is written in response and time.

[0132] The display data reading unit 15 sequentially reads the terminal-side generated one-day display data Sfi (FIG. 7(b)) of the terminal-side generated one-day display list SFi into the display data memory 14.

[0133] The time task code determination unit 13 preferentially reads the time box of the terminal-side generated one-day presentation data Sfi read into the presentation data memory 14, and if no abnormal data (the column is empty) is written, it determines that the data is normal and starts the normal TASK "0" processing unit 12. Furthermore, if abnormal data (negative value: -1 or -2) is written, the terminal side skip unit 16 (terminal side skip task) is activated to skip the one-day presentation data Sfi generated for the terminal side, and the TASK "0" processing unit 12 (TASK "0") is activated again.

[0134] When the performance evaluation value Gmi for the content displayed in the response box of the data Shi for obtaining the individual evaluation score is set to the highest score ma, the terminal side skip unit 16 (terminal side skip task (TASK "1")) writes (-1) in the time box of all subsequent one-day presentation data Sfi generated for the terminal side that have the same serial code (content) in the one-day presentation list SFi generated for the terminal side. As a result, the content of (-1) is skipped and removed from the display (not displayed).

[0135] Furthermore, if (-2) has already been written in the time box of the terminal-side generated one-day presentation data Sfi, the terminal-side skip unit 16 (terminal-side skip task (TASK "1")) causes the processing of terminal-side TASK "2" to stop the display for TASK "0" and specify the next presentation data.

[0136] As a result, in the learner terminal 10i, the task code of the terminal-side generated one-day presentation data Sfi is set to time (TASK←time), and the content of (-2) is skipped and removed from the display (not displayed). In other words, the process of task code TASK←TASKE "2" is executed in the learner terminal 10i (see FIG. 22).

[0137] In other words, by setting time to "-", the process of TASK [1] or TASK [2] can be executed in addition to TASK "0" without changing the application on the learner's terminal. That is, (A1). The TASK process (TASK “0” process) is composed of multiple tasks that execute without separating how to display learning content and how to acquire data according to the set task code; (A2) A skip task process for skipping the designated presentation data and designating the next presentation data ((TASK “1”), (TASK “2”) on the terminal side), (A3) A process of reading the column for obtaining reaction time with priority, and if this column is "empty", setting the task code to TASK processing and executing it, and if a negative value ("-1" or "-2") is written, setting a skip code ((the code for TASK "1"), (the code for TASK "2" on the terminal side)) to execute skip task processing; (A4) The combination of the serial code of the presentation data in the presentation list, the individual performance evaluation value in the column for obtaining the individual performance evaluation value, and the reaction time in the column for obtaining the reaction time is treated as reaction data, and a process is performed in which this reaction data is sent to the analysis center as a reaction list in response to a transmission instruction.

[0138] In addition, data Shi for obtaining one's own individual evaluation score value that does not have "-1: learning terminal side determined display skip coefficient αi" written in the time field is called normal data shai for obtaining one's own individual evaluation score value, and data shbi for obtaining one's own individual evaluation score value for skipping that has (-1) written in the time field (see Figure 11).

[0139] <Overall explanation> The system configured as above will be explained using a sequence diagram. In this embodiment, the web side (wb) and the local side (Wc) are shown, and the generation of the presentation list will be explained first. Also, the learner terminal 10i will be explained as one unit, and will be simply referred to as the learner terminal 10.

[0140] As shown in Fig. 12, the person in charge on the local side (Wc) operates the person in charge terminal (not shown) to input the name of the contracting school (which may be an individual), age, school name (also called group name), gender, learning course (e.g., English or Japanese for first-year junior high school students), learning level (beginner, intermediate, advanced), contract details (learning schedule, fee), name of the instructor, learner terminal address, email address, telephone number, registration date and time, etc. (d10), and performs various information registration processing (d11). In this registration processing, the various registration processing units on the local side display input screens on the person in charge terminal (not shown) to allow input.

[0141] The various information registration process stores these as learner information Ai (A1, A2, . . . A1, Ab2, . . . ) in the local DB server 500 (d12). In addition, the person in charge at the local side Wc decides what level of difficulty and how many pieces of content the learner should do per day, and how long the intervals should be, in order to achieve the desired results based on the difficulty of the course, learning level, etc., based on the learner information Ai.

[0142] In this case, difficult words are assigned from the first day until a predetermined date, and easy words are assigned from the predetermined date (for example, the 10th) onwards, and the number of repetitions is determined to be one from the first day until, for example, the 5th. That is, the schedule conditions SJ for the interval in units of one day and the schedule SJ conditions for the interval of 10 days are input to the local server 400 from the person in charge terminal (not shown) (specifically, the identification code of the schedule conditions). That is, the experiential schedule conditions shown in Figure 4(b) are generated and stored in memory (not shown).

[0143] Furthermore, the person in charge may decide what content to study, how many times per day, and at what intervals, and store this as schedule conditions (programmed in Excel) in a memory (not shown). A plurality of schedule conditions are generated in advance according to the level of difficulty. Furthermore, the schedule table generating unit 410 generates a schedule TB based on empirical schedule conditions (also simply referred to as schedule conditions) and stores the generated schedule TB in a schedule TB memory 510 .

[0144] In this state, the local-side DB server 500 sends the learner basic identification information Bi [including learner-specific identification information SGi: name, group name, difficulty level N, account, registration number, learner terminal address, email address, and telephone number)] to the web-side DB server 300 and stores (copies) it (d14).

[0145] In addition, the local side presentation list generation unit 420 of the local server 400 sequentially designates the learner information Ai, and generates a local side presentation list TRi (TR1, TR2, ...: see Figure 6(a)) for the learner information Ai for each learner over the analysis period (e.g., 60 weeks), which is the learning period, and stores it in the local side DB server 500 (d20: local side presentation list generation process).

[0146] [Process for generating local presentation list TRi (d20)] The local side presentation list generating process (d20) of the local side presentation list generating unit 420 generates the local side presentation list TRi based on the difficulty level of each course, learning level, and the like.

[0147] The person in charge decides what level of difficulty and how many pieces of content should be given to the learner per day, how long the intervals should be, etc., in order to achieve the desired grades based on the learner information Ai and the difficulty level of the course and learning level. In this case, difficult words are assigned from the first day until a specified date, and easy words are assigned from the specified date (for example, the 10th) onwards, and learning content A (for example, desk) is repeated once from the first day until, for example, the 5th. In other words, the schedule conditions SJ for the one-day interval and the schedule SJ conditions for the 10-day interval are entered into the local server 400 from the person in charge's terminal (specifically, the identification code of the schedule conditions). A plurality of schedule conditions are generated in advance according to the level of difficulty.

[0148] As a result, the schedule table generation unit 410 generates in memory (not shown) the schedule conditions of the learner for the intervals of these conditions (the person in charge may decide what content to study, how many times, and at what intervals per day, and store this as schedule conditions (for example, an Excel program) in memory (not shown). In other words, the experiential schedule conditions shown in Figure 4(b) are generated.

[0149] Then, the local presentation list generation unit 420 sequentially reads the schedule data SDi (SD1, SD2, ...: see Figure 5) from the schedule TB memory 510 (d23), and each time it reads this schedule data SDi, it reads (d21) one of the contents (words) contained therein that corresponds to the difficulty level N, etc., randomly from the content DB (see Figure 3(a)) in the content DB memory 520. (For example, desk (number 111))

[0150] Then, a local side presentation list TRi (for example, 60 weeks' worth) as shown in FIG. 6(a) is generated in the presentation list memory 530 (d25).

[0151] Although not shown, the local-side presentation data Tri included in the local-side presentation list TRi is associated with learner-specific identification information SGi, experience schedule conditions (specifically, an identification code), and the date and time of generation. A collection of the local-side presentation data Tri is the local-side presentation list TRi. Then, the web side presentation list sending unit (not shown) of the local server 400 generates (d22) a web side presentation list WRi [WR1, WR2, ...: serial, word (no answer), response, time: (see FIG. 6(b))] and sends (d24) it to the web side (Wb) via the LAN 80. In practice, it is preferable to arrange the time next to the serial.

[0152] Each web-side presentation data Wri in the web-side presentation list WRi is associated with learner identification information SGi, experience schedule conditions, creation date and time, etc.

[0153] (Local web presentation list generation process) The above-mentioned local web presentation list generation process will now be explained in more detail. The local Web presentation list generation process sequentially allocates (specifies) records in the local presentation list TRi (see FIG. 6(a)).

[0154] Then, the information including the serial code contained in the local presentation data Tri (see Figure 6(a)) of each record specified, the time box [(specifically, a code indicating a box where the time can be written is associated, and it is also called the time column)], the word (for example, desk (no answer: word code 111), the response box (specifically, a code indicating a box where the results can be written), MD (date), the TASK "0" code, etc.) is extracted as web presentation data Wri (see Figure 6(b)).

[0155] The aforementioned TASK "0" code means the task code of the normal process (TASK "0") to be performed when the time box (column) is "empty". For example, processes such as D (drill), T (test), O (survey), etc. (code that specifies how the content is displayed and what kind of response is obtained). To show this, in Figure 5, the column B-R01(K)... is simply written as TASK "0" to indicate that it is a task code for normal processing (display control program).

[0156] The local Web side presentation list generation process then converts a set of Web side presentation data Wri (for 60 weeks) into a file as a Web side presentation list WRi and transmits it to the Web side (Wb). The web presentation data Wri of the web presentation list WRi has learner identification information SGi and the like added thereto.

[0157] That is, a web presentation list WRi is generated for each learner identification information SGi for each course on the local side and transmitted to the web side (wb). The web-side DB server 300 on the web side (Wb) performs a web-side presentation list collection process (d27) to store the web-side presentation list WRi [serial code, time box, word (no answer), response], response box, MD (year / month / date), task code: see Figure 6(b)] in memory each time it receives it (d26), and then performs a process to generate a presentation list to be sent to the learner (d28).

[0158] (Processing for generating presentation list to be sent to learners) The web-side presentation list collection process (d28) of the web-side DB server 300 determines whether a web-side preparation presentation list SRi (Figure 7(a)) having learner-specific identification information SGi of this web-side presentation list WRi has been generated in memory each time it receives the web-side presentation list WRi from the local side.

[0159] If not generated, a memory (area) for storing this web side preparation presentation list SRi (FIG. 7(a)) is generated. If it is not stored, the web side presentation data Wri of the web side presentation list WRi is sequentially specified from the local side, and for each specification, a serial code (e.g., 111) is read (copied), and this is associated with the time box, response box, MD (year, month, day), and TASK “0” code (e.g., D) (web side preparation presentation data Sri (no content: see Figure 7(a))).

[0160] In other words, the presentation list generates presentation data that is a set of the serial code of the learning content in the content list, a column for obtaining the individual evaluation score, a column for obtaining the reaction time, and a task code (D or T, etc.) for the TASK processing, arranged for the same learning content and for different learning contents.

[0161] Then, it is stored in the web-side DB server 300 (d29). Although not shown, the web-side preparatory presentation data Sri is associated with learner identification information SGi and the like. This set of web-side preparation and presentation data Sri (for 60 weeks) is called a web-side preparation and presentation list SRi. In other words, a web-side preparation and presentation list SRi is generated for each learner.

[0162] The web-side preparation presentation list SRi (questions, no answers: 60 weeks worth) is stored for each learner's identification information SGji (SGa1, SGa2, . . . , SGb1, SGb2, . . . ). Next, a description will be given with reference to the sequence diagram of Fig. 13. As shown in Fig. 13, the content extractor 430 of the local server 400 performs processing to extract content from the local-side DB server 400 (d30).

[0163] (Content retrieval process (d30)) Next, the content retrieval process in the local DB server 500 will be described. The content extraction process sequentially specifies the local presentation list TRi [(TR1, TR2, . . . : course, difficulty level, task code, learner identification information SGi: 60 weeks worth) stored in the local DB server 500.

[0164] Then, as shown in FIG. 13, for each designation, the local side presentation data Tri (see FIG. 6(a)) in the local side presentation list TRi is designated in sequence (d32). For each such specification, the content having the serial code contained therein is searched for in the content database, and the searched content is extracted (copied) as actual content Cmi [(for example, the English word (desk) and the answer (desk), the English word (Apple) and the answer (apple), ...)] (d34). Extraction is performed for 60 cycles (i.e., 120 days' worth).

[0165] These contents are called real contents Cmi (serial, English words, answers, schedule conditions (identification code), course, etc.), and are sent as a file to the web-side DB server 300 (d36).

[0166] Meanwhile, the Web-side DB server 300 receives the actual content Cmi (Cma, Cmi, . . .: English word, answer, serial, schedule condition, course, . . .) and stores it in the Web-side content memory 310 (d38).

[0167] Then, the web-side presentation list generating unit 220 of the web-side (wb) application server 200 generates a presentation file GRi (GRi, GRi, . . . : also called a sending list) to be sent to the learner by a process of generating a presentation list to be sent to the learner (d40).

[0168] (Processing for generating presentation list to be sent to learners) The process of generating a presentation list to be sent to learners sequentially specifies the learner identification information SGi (SG1, SG2, . . . ) of the web DB server 300 (d42). Then, for each designation, a web-side preparation presentation list SRi (see FIG. 7(a)) having the designated learner-specific identification information SGi is designated.

[0169] Then, for each designation, the web side preparation and presentation data Sri included in the web side preparation and presentation list SRi (FIG. 7(a)) is designated in sequence (d46). Then, for each designation, the local side presentation data Tri (see FIG. 6(a)) having the serial (serial code) number included in the designated web side preparation presentation list SRi is retrieved (d48).

[0170] Then, the process of generating a presentation list to be sent to a learner reads the serial (for example, 111) included in the allocated local side presentation data Tri (record: FIG. 6(a)). Then, the actual contents Cmi (Cmi, Cmi, . . .: English word (e.g., desk), answer (e.g., desk), schedule condition, course, . . .) of the words of the read serial codes are read from the contents DB (d50).

[0171] Then, the read actual contents Cmi are associated with the designated web-side prepared presentation data Sri. The aforementioned web-side prepared presentation data Sri may have the serial (e.g., "111") and response boxes empty (e.g., nothing written) and the display skip coefficient αi "-2: negative value" written in the time box, as determined by the latent memory analysis process described below. If the display skip coefficient αi "-2: negative value" is written in the time box, it is determined that a negative value has been written on the terminal side. Then, the task code is set to TASK←time on the terminal side. In other words, time is given priority.

[0172] Hereinafter, the 60 weeks' worth of web-based daily preparation presentation list SRi and the 60 weeks' worth of actual content Cmi are collectively referred to as a presentation file GRi (including task codes (not shown): FIG. 14) to be sent to learners. The above processing is the preparation before the learner accesses the site. Next, the processing when the learner actually accesses the site will be explained using the sequence diagram in Figure 15. However, it is explained that learners are given a medium paper with a QR code (registered trademark) to access the center.

[0173] As shown in Figure 15, the learner reads the QR code (registered trademark) with the learner terminal 10 (d51), accesses the web side (wb) (terminal number, URL, date and time, etc.), and performs the download process for the learner terminal side application (d50). The browser displays the application icon on the screen, and selecting this application icon displays the center side URL, and selecting this URL downloads TASK "0", skip tasks (TASK "1", TASK "2", and related programs) from the center side.

[0174] In this download, the web server 240 receives the learner terminal number, date, time, etc. and outputs them to the application server 200, which then reads the learner terminal side application (also called the terminal side application) from the memory (d52) and sends it to the learner terminal 10 (d50). This terminal-side application is responsible for tasks such as TASK "0" (display control program), TASK "1" (first terminal-side display skip task), TASK "2" (second terminal-side skip task), basic screen display, sending and receiving, etc., and is associated with accounts, expiration dates, etc.

[0175] When downloading the terminal-side application, the web server 240, upon access from the learner terminal 10, displays an input screen on the learner terminal 10 for inputting the above learner information Ai (group ID, ID code, password, name, age, grade, school name, email address, mobile phone number, etc.). The web server 240 generates a registration number upon receiving the learner information Ai from the learner terminal 10, and stores the registration date and time (internal timer) in association with the number. The web server 240 also sends the learner information Ai, the registration date and time, and the registration number to the local DB server 500, where they are stored and managed in memory.

[0176] The learner terminal 10 downloads the learner terminal-side application to its memory (d54). This learner terminal-side application will be described later. The learner terminal 10 performs initial setting processing such as displaying the application icon of this learner terminal-side application on the screen (d55).

[0177] When this application icon (not shown) is selected, the application on the learner terminal 10 displays a login screen (FIG. 16(a)) including the group ID, ID code, password, registration number, etc., and transmits these as login information to the web server 240. The web server 240 then authenticates (verifies) the account in the login information (d57) and transmits the authentication result to the learner terminal (d56). The learner terminal 10 displays, for example, an input screen for course selection, level, etc. (see Fig. 16(b) and Fig. 16(c)). Also, as shown in Fig. 17, the learner's level, the number of items remaining until complete acquisition, and a button to return to course selection are displayed. The number of items remaining is transmitted from the center side.

[0178] Then, the terminal-side application transmits the input information (learning request information GYi) to the web side (d56). Every time the web server 240 receives learning request information GYi [group ID, ID code, password, registration number, difficulty level (course selection, level), date and time], it compares it with the learner identification information SGi in the web-side DB server 300, and if the comparison is successful, it performs the learner presentation list transmission process (d60).

[0179] (Student presentation list sending process (d60)) The web server 240 retrieves the presentation file GRi to be sent to the learner [actual content Cmi and actual content Cmi [(content number, English word (e.g., desk), answer (e.g., desk), schedule condition (including task code), course, ...) Figure 14] stored in the memory of the web-side DB server 300 (d62).

[0180] Then, the presentation file GRi [SRi (srai, srbi), (including task code), Cmi] to be sent to the learner is retrieved, which has the learner identification information SGi (learner identification information Bi, learning terminal information Di, etc.) included in the learning request information GYi.

[0181] Then, all the web side preparation and presentation data Sri of the web side preparation and presentation list SRi (for 60 weeks) included therein are read out (d63). Furthermore, the actual content Cmi (e.g., desk) having the serial included in the web-side preparation presentation data Sri of the terminal-side generated one-day presentation list SFi (FIG. 7(b)) is read (d65). That is, one day's worth of actual content Cmi is read.

[0182] Then, this one day's worth of actual content Cmi is put into a file (hereinafter referred to as one day's content file CFi or content list), and the one day's generated presentation list SFi for the terminal side [serial code, time box, response box, MD (year / month / date), task code: one day] is combined with it (one day's learner sending file GFRi) and sent to the learner terminal 10 (d64).

[0183] The web-side presentation list generation unit 220 associates the web-side preparation presentation data Sri of the terminal-side generated one-day presentation list SFi (FIG. 7(b)) with the task code (D or T, . . .) of the display control program that processes TASK (0:D or K, T, . . .). It also associates the code of the skip display control program that processes TASK "2".

[0184] The learner terminal 10 receives this one-day learner delivery file GFRi (one-day terminal-side generated presentation list SFi, one-day learner-specific content file CFi) and overwrites and saves it in the download memory (not shown) (d66, d68).

[0185] This means that the learner terminal 10 is now ready to start learning. That is, the web-side application server 200 sequentially generates a terminal-side generated one-day presentation list SFi (including task codes), which is a web-side preparation presentation list SRi for each day, based on the presentation file GRi to be sent to the learner, and transmits each generated one-day presentation list SFi (including task codes) to the learner terminal 10 together with a learner-specific one-day content file CFi (content list) (see Figure 14).

[0186] Next, the processing after the start of learning will be described with reference to FIG. (Start learning) As shown in FIG. 15, the learner operates the learner terminal 10 to select an application icon (not shown) on the screen and start learning (d70). The learner application of the learner terminal performs a learner-side presentation preparation process (d70) when an application icon (not shown) is selected (learning start).

[0187] (Learner's side presentation preparation process on the learner's terminal (d70)) The learner's side presentation preparation process, which is a terminal side application of the learner's terminal, will be described with reference to the flowcharts of FIGS. As shown in Figure 18, the learner terminal application of the learner terminal 10 has a learner side presentation preparation process, which, in conjunction with the downloading of a one-day learner delivery file GFRi, generates in memory a personal assessment response list SHi (see Figure 7(c)) similar to the one-day presentation list SFi for terminal use contained in the file (S9). Note that "serial" may be a serial code for the same learning content or a serial code for a different learning content.

[0188] As shown in Fig. 7(c), this individual evaluation response list SHi is a record group (MD and SGi are associated but not shown) consisting of serial (e.g., "101"), response, time, task code, etc. In other words, a table of a record (column) group with the same number as the terminal-side generated one-day presentation list SFi (Fig. 7(b)) is generated. The items are TASK, serial, response, time, etc., and each record is associated with a record number. Next, the presentation data writing unit 15 sequentially specifies (specifies record numbers of) the one-day presentation data Sfi generated for the terminal side in the one-day presentation list SFi (Fig. 7(b)) generated for the terminal side included in the one-day learner delivery file GFRi [one-day presentation list SFi generated for the terminal side (Fig. 7(b)), one-day content file CFi for each learner] downloaded to the download memory, and reads out the specified one-day presentation data Sfi into the presentation data memory 14 (S10).

[0189] The terminal-side generated one-day presentation data Sfi (MD and SGi are associated but not shown) consists of web-side pre-latent memory assessment presentation data Sfai (Figure 7(b)), or first latent memory assessment data Sfbi (Figure 7(b)), or second learner-specific content-specific latent memory assessment data Sfci, as shown in Figure 7(b).

[0190] The web-side pre-latent memory assessment presentation data Sfai includes records such as serial (e.g., "111"), a response box (□: empty), a time box (□: empty), and a task code, as shown in Figure 7(b).

[0191] The second potential memory assessment data Sfci (Figure 7(b)) refers to a record consisting of serial (e.g., "111"), a response box (□), a time box, and (web-side decision display skip coefficient βi(-2) (task code of TASK "2")), etc., as shown in Figure 7(b).

[0192] The first latent memory level determined data Sfbi and the second latent memory level determined data Sfci are collectively referred to as learner-specific content-specific latent memory level determined data Svi. Then, the time task code determination unit 13 (terminal side skip task of the learner terminal side application) reads the time box of the one-day presentation data Sfi generated for the terminal side that was read into the presentation data memory 14 in step S10 with priority, and determines whether or not abnormal data ("-": negative value) has been written (S11). If no abnormal data (the column is blank) is written in the time box, it is determined to be normal, and the TASK "0" processing unit 12, which is a normal task [(TASK "0" (time is blank) processing: display control program)], is started to execute the following display processing. For example, it executes processing D.

[0193] If this normal task (TASK "0") is normal, the time task code determination unit 13 (the application on the learner terminal side) sets the code of TASK "0" (time is empty) in the TASK code setting box (not shown) of the TASK "0" processing unit 12 (TASK←(0)). Since time is "empty," for example, "D" of the terminal-side generated one-day presentation data Sfi designated in step S10 is set in the TASK code setting box (job queue: not shown). As a result, the TASK "0" processing unit 12 reads the English word (e.g., desk) having the serial number and the answer (desk) contained in the terminal-side generated one-day presentation data Sfi (web-side pre-latent memory assessment presentation data Sfai (Figure 7(b))) from the learner-specific one-day content file CFi [content number of one day's content, English word, answer, serial number, schedule conditions, course, ...: content list] (S14).

[0194] For example, in the case of display control program D, if serial is "111" (for example, the word "desk"), the English word (desk) is displayed, and when the answer display button is selected, the normal display processing is performed, such as displaying the answer (desk) and the individual evaluation item (not good at all, not good, a little better, good) (S16: see Figure 20).

[0195] A specific example of the display process in the learner terminal 10 is shown in FIG. As shown in Figure 20(a), for example, Emphasis is displayed in the question content field ga of the actual content Cmi corresponding to the serial included in the one-day presentation data Sfi generated for the terminal side, and a response measurement bar gb is displayed, and further, an answer display button gc is displayed. Preferably, after thinking about it for a few seconds, the answer display button gc is selected. The display of the response measurement bar gb indicates the time measured by an internal timer (not shown), and is the time measured from the time when, for example, Emphasis is displayed in the question content field ga to the time when one of the personal evaluation items (not good, bad, a little better, good) in Figure 20(b) is selected.

[0196] By selecting the answer display button gc, a personal assessment screen for the level of mastery of the content (words) shown in FIG. 20(b) is displayed (personal assessment items: not good at all, not good, a little better, good). When selecting the grade item Fm, you should compare the displayed meaning with your implicit memory (however, if you are completely unable to remember a word, do not try to remember it on the spot; just look at the answer and move on to the next word). The display process may be performed as shown in FIG. 21(a) and FIG. 21(b). Then, a reaction result writing process (not shown) included in the TASK "0" processing unit 12 performs a process of writing the reaction.

[0197] As shown in Figure 18, the TASK "0" processing unit 12 specifies a record in the personal evaluation response list SHi (see Figures 7(c), 10, and 11) having the specified record number in the one-day presentation list SFi generated for the terminal side, and writes the response in the corresponding item of this record. Specifically, the grade rating item Fm selected by the learner is written in the response box. This selected grade rating item Fm is written as the grade rating value Gmi (also called the learner content rating value), and the reaction time tm (measured time) when the grade rating item Fm was selected is written in the time box (S18). The grade rating Gmi (student content rating) is specifically "0" for completely bad, "1" for bad, "2" for slightly better, and "3: maximum score ma" for good.

[0198] Then, the process proceeds to step S20 shown in FIG. TASK "1" included in the terminal side skip unit 16 monitors the personal evaluation response list SHi, and as shown in Figure 19, specifies a record in the personal evaluation response list SHi and determines whether the performance evaluation value Gm written in the response box of the personal evaluation score value acquisition data Shi of the specified record is the highest score ma ("3") (S20).

[0199] In step S20, if the terminal-side skip unit 16 (TASK "1") determines that the highest score is ma ("3"), the terminal-side skip unit 16 (TASK "1") writes the learning terminal-side determined display skip coefficient αi "-1 (also called skip code)" into the time box of all terminal-side generated one-day presentation data Sfi (web-side pre-latent memory judgment presentation data Sfai (Figure 7(b))) that have the same serial (e.g., 111:desk) after the specified sfi in the terminal-side generated one-day presentation list SFi (Figure 7(b): presentation list for one learning session) specified by the terminal-side skip unit 16 (TASK "1") (S22).

[0200] That is, the first latent memory degree judged data Sfbi (FIG. 7(b)) is generated in the terminal-side generated one-day presentation list SFi. Therefore, on the learner terminal 10 side, all serial (content) items after the one with the highest score in the presentation list (SFi) for one day will be set to (-1) and will be skipped (not displayed, removed from display). In other words, the task code is set to TASK←time(-1), and each time TASK "1" specifies the first latent memory assessment data Sfbi in the one-day presentation list (SFi), time is judged to be a negative value and the display is skipped (not displayed, removed from display). If it is determined in step S20 that the score is not the highest, the normal display control program (TASK "0") writes the grade evaluation item Fm ("0" or "1" or "2") into the response (S21).

[0201] Then, the terminal side skip unit 16 (TASK "1") retrieves the personal evaluation response list SHi (Figure 7(c)), as shown in Figure 19, retrieves the data Shi for obtaining the personal evaluation score value having the serial (e.g., 111:desk) of the specified one-day presentation data Sfi generated for the terminal side, and writes "3" to the response of this data Shi for obtaining the personal evaluation score value (S24).

[0202] Next, the display control program (TASK "0") determines whether or not there is any other one-day display data Sfi generated for the terminal side in the one-day display list SFi generated for the terminal side (FIG. 7(c)) (S26). In step S26, if TASK "0" determines that there is other one-day display data Sfi generated for the terminal side in the one-day display list SFi generated for the terminal side (Figure 7(c)) (yes), the data is updated to the next one-day display data Sfi generated for the terminal side, and the process returns to step S10 in Figure 17 (S36).

[0203] Furthermore, if TASK "0" determines in step S26 that there is no other one-day presentation data Sfi generated for the terminal side in the one-day presentation list SFi generated for the terminal side, it performs a process of transmitting the set of data Shi for obtaining individual evaluation scores in the memory as the learner response list Hi by a transmission process (at the time the send button is selected) that transmits the response list Hi to the web side (S38).

[0204] On the other hand, in step S11 shown in FIG. 18, if the time task code determination unit 13 (included in the display control program) determines that abnormal data (negative value: "-1 or -2") has been written in the time of the one-day presentation data Sfi generated for the terminal side in the presentation data memory 14, the terminal side skip unit 16 (TASK "1") is activated, and this terminal side skip unit 16 (TASK "1") determines whether the time is (-1) (S28).

[0205] In step S28, if the terminal side skip unit 16 (TASK "1") determines that (-1) has been written in the time box, it determines to remove the display of the specified one-day generated terminal side presentation data Sfi (first latent memory level determined data Sfbi) (also called skip: do not display and display the next sfi), writes "3" (no writing is required) in the response box of the data Shi for obtaining the individual evaluation score value, which has the serial of the specified one-day generated terminal side presentation data Sfi (first latent memory level determined data Sfbi), and moves the process to step S26 in Figure 19 (S30).

[0206] In other words, if (-1) is written in the time of the specified one-day presentation data Sfi (first potential memory assessment data Sfbi) generated for the terminal side (i.e., "-"), this Sfi will be skipped and the next Sfi will be specified, so the content of this specified one-day presentation data Sfi generated for the terminal side will not be displayed.

[0207] Also, in step S28, if the terminal side skip unit 16 determines that the time of Sfi is not (-1) but "-2: web side determined display skip coefficient βi", it writes "3" (or does not have to) to the response of the data Shi for obtaining the personal evaluation score value, and writes (-2) to the time, as in the process of step S30, and moves the process to step S26 in Fig. 19 (S34). This process is called the process of TASK "2".

[0208] That is, when the time of the designated terminal-side generated one-day presentation data Sfi is "-2: web-side determined display skip coefficient βi", the display is skipped as in the case of (-1).

[0209] That is, if abnormal data has been written, it is determined whether (-1) or (-2) has been written to time, for example.

[0210] If an abnormality is determined, the TASK "0" processing unit 12 outputs information (Sfi=Sfi+1) to specify the next one-day presentation data Sfi for the terminal side after the specified one-day presentation data Sfi for the terminal side.

[0211] The writing of (-2) into the terminal-side generated one-day presentation data Sfi is a web-side process, so the web-side process will be described below.

[0212] That is, as shown in FIG. 22, when the time is empty (TAKE "0"), the learner terminal 10 performs the normal display control program (TASK "0") processing, for example, displaying apple, prompting the learner to press a button, displaying the answer, and giving a grade; when the time is "-1" or "-2", the terminal-side skip unit 16 skips the display (FIG. 22 is an example in which the grade value is written as "3").

[0213] Then, as shown in Figure 15, a display processing unit (not shown) determines whether the send button has been selected (d72), and if selected, transmits the set of reaction data hi stored in memory up to the current point in time to the web side as a reaction list (d74). The web side stores this reaction list Hi in memory (d78). Next, the processing on the web side will be explained.

[0214] (Web side processing) The web-side processing will be explained using the sequence diagram of Fig. 23. The potential memory level confirmation unit 210 of the application server 210 reads the response list Hi from the memory and performs web-side content performance value definition processing (d90), as shown in Fig. 23.

[0215] (Web-side content performance value definition processing (d90)) The web-side content performance value definition process will be described using the flowchart in Fig. 24. As shown in Fig. 24, the web-side content performance value definition process specifies reaction data hi (including task code) in reaction list Hi stored in memory from the learner terminal 10 (S40).

[0216] If the task code included in this response data hi is the TASK←TASK "0" code, the serial (content) included therein is read (S42), and it is determined whether or not a coordinate system Vci (specifically, a function: see Figure 9) for confirming latent memory level that has this serial (content) has been generated (S44).

[0217] If not generated, a coordinate system Vci for checking latent memory level having this serial (content) is generated in memory (S46).

[0218] Then, all the response data hi having serial (e.g., desk) included in the response list Hi (for one day) are read (specified), and the average of the performance rating values ​​Gmi (e.g., "0," "1," "2," or "3") of all the responses included in this response data hi is calculated as the individual performance average rating value Pm per learning session (S48).

[0219] Furthermore, without averaging, all reaction data hi having serial (e.g., desk) included in the reaction list Hi (for one day) can be read (specified), and if the response time of any of this reaction data hi contains "-1", the first highest score "3" can be written as the performance evaluation value Gmi ("3") (Time is, for example, "1.5 seconds").

[0220] Then, the individual's average performance score Pm (e.g., "1") of the response included in the specified response data hi [serial, time (e.g., 1.5 seconds), response (e.g., "1"), ... learner identification information SGji, number of presentation cycles Xi (1st day, 2nd day, ...)] is read (S50).

[0221] Then, the individual's own average grade Pm (e.g., "1") is defined in the Xi portion of the coordinate system Vci for confirming latent memory corresponding to the number of presentation cycles Xi (learning times) included in the response data hi [serial, time (e.g., 1.5 seconds), response (individual's own average grade Pm: e.g., "1"), task...learner-specific identification information SGji, number of presentation cycles Xi (1st day, 2nd day,...)] (S52: see Figure 9).

[0222] Then, it is determined whether there is other reaction data hi in the reaction list Hi (for one day) of the learning session, and if there is other reaction data hi, the reaction data hi is updated to the next and the process returns to step S42 (S54).

[0223] Also, in step S44, if it is determined that the coordinate system Vci (specifically, a function) for confirming the latent memory degree having the serial (content) included in the reaction data hi has already been generated in the memory, the process returns to step S50. Therefore, as shown in FIG. 25, the individual grade average Pm is sequentially defined in the coordinate system Vci (specifically, a function) for confirming the latent memory level.

[0224] In FIG. 25, the horizontal axis shows the number of presentation cycles Xi as the 1st, 10th, 11th, 12th, 13th, and 14th days (serial: 155, temple). In this example, the web-side generated one-day presentation list SFi for each of the 1st to 10th days has one terminal-side generated one-day presentation data Sfi (presentation condition is 1 time) having, for example, serial (155:temple), the web-side generated one-day presentation list SFi for the 11th day has two terminal-side generated one-day presentation data Sfi (presentation condition is 2 times: number of repetitions is 2), the web-side generated one-day presentation list SFi for the 12th day has one terminal-side generated one-day presentation data Sfi (presentation count is 1 time), the 13th day has three presentations, and the 14th day has two presentations, which are sent from the web side to the learner terminal 10, and these reactions are received on the web side as a reaction list Hi.

[0225] After this web-side content performance value definition process (d90), the latent memory confirmation unit 210 specifies the number of presentation cycles Xi on the horizontal axis of the coordinate system Vci (specifically, a function) for latent memory confirmation (d92), as shown in Fig. 23. Then, a regression line generation process (d94) is performed.

[0226] (Regression line generation process (d94)) The regression line generation process of the latent memory level checking unit 210 will be described with reference to Fig. 25. This latent memory level checking coordinate system Vci is generated for each learner identification information SGi and for each Serial (content).

[0227] For example, as shown in FIG. 25, they are defined as P1, . . . , P10, P11, P12, P13, and P14. For example, serial(155:temle) indicates a case where the evaluation for the presentation on the first day (one-day presentation list SF1 generated for terminal side) was completely poor (personal grade average Pm (average) is "0"), but the evaluation for the presentation on the 10th day (one-day presentation list SF10 generated for terminal side) was good (personal grade average Pm (average) is "3"). Also, the evaluation for the presentation on the 11th day (one-day presentation list SF11 generated for terminal side) was "a little better: personal grade average Pm (average) is "2"), and the evaluations for the presentation on each of the presentation lists (SF12, SF13, SF14) on the 12th, 13th, and 14th days were all good (personal grade average Pm (average) is "3").

[0228] In reality, the interval for simple words is long, so it is once every 10 days (1, 2, 3 or 4 times; most often once). Each time response data hi is specified, the latent memory confirmation unit 210 reads the individual performance average rating value Pm at the presentation cycle number Xi on the horizontal axis of the coordinate system Vci (specifically, a function) for confirming latent memory corresponding to the presentation event cycle number included in this response data hi, and the individual performance average rating value Pm (Pm = Pm-1, Pm-2, ...) defined for the previous presentation cycle number Xi, and sequentially calculates the regression line yi (yi = aXi + b) and defines it in the coordinate system Vci (specifically, a function) for confirming latent memory as shown in Figure 25.

[0229] Then, it is determined whether this regression line yi (yi=axi+b) reaches the highest dotted line MFi (reaches the highest point ma ("3")) (d96). The highest point ma may have a certain range, from 2.8 to 3.0. Preferably, it is 3.0. If it is determined that the regression line yi (yi = axi + b) has reached the highest dotted line MFi (reached the highest point ma ("3")), it is determined whether the individual's average performance rating Pm is defined at the highest point ma ("3") in the coordinate system Vci for confirming latent memory level from three cycles before (including the specified presentation cycle Xi) the specified presentation cycle Xi (event presentation count) consecutively up to the specified presentation cycle Xi (d98).

[0230] If it is determined that the individual's grade point average Pm is the highest point ma ("3") consecutively and the regression line yi (yi=axi+b) reaches the highest dotted line MFi (reaches the highest point ma ("3")) (the two conditions are met), it is determined that the content is implicitly memorized (d100). By this processing, as shown in Figure 25, for example, the evaluation of the presentation list (1-day presentation list SF10 generated for terminal side) on the 10th day is good (personal performance average evaluation value Pm (average) is "3"), but the regression line y10 on the 10th day has not reached the highest dotted line MFi, so it can be determined that the content has not yet been implicitly memorized. In other words, it is determined that this content is not yet familiar (not internalized). However, when the regression line y14 on the 14th day is calculated, this regression line y14 on the 14th day reaches the highest dotted line MFi, and the individual grade point average Pm for P12, P13, and P14 consecutively reaches the highest score ma ("3"), so the learner is judged to have mastered this content (has sufficient implicit memory).

[0231] That is, even if the content presented in the presentation list (1-day presentation list SFi generated for the terminal side) is mistakenly selected as "good," or even if a lot of content is studied the day before or several days before the test, the influence of this can be eliminated. Also, even if "good" is mistakenly selected on the first day, the influence of this can be eliminated. It is extremely rare for the performance rating to be the highest three times in a row (study times).

[0232] Also, if a response list Hi is received for the first time, containing response data hi for which the performance evaluation value Gmi is the highest for a word that first appeared on the first to third days (first day being preferable), the response contained in that response data hi may be forcibly set to, for example, (1) and defined in the performance definition coordinate system HZi.

[0233] If it is determined that the learner has mastered this content (has implicit memory), all subsequent local presentation data Tri in the local presentation list TRi (for 60 weeks) that have the serial (e.g., temple) of the specified response data hi are specified, and (-2) is written to time (d102).

[0234] Then, as shown in FIG. 23, it is determined whether or not another response list hi exists in the memory (d104). If it does not exist, the reaction list Hi is sent to the local server (d106). In this case, the response data hi that meets the first and second conditions and has reached a high level of latent memory may be associated with a high latent memory flag (third condition), since the learner will no longer need to study this content.

[0235] The local server associates this reaction list Hi with TRi and stores them (see FIG. 26) (d108). This allows another process (not shown) to analyze this reaction list Hi, making it possible to remove the contents of the serial code of the reaction data hi for which a potential memory high flag exists from the experience schedule conditions and schedule table, and to allocate other contents.

[0236] Furthermore, each time there is a latent memory level flag (third condition), the total number of presentations is subtracted from the experience schedule condition of the learner, and the result is transmitted to the learner terminal 10 as the remaining number at each login (see FIG. 17). A specific example of a regression line for each individual content is shown in a table in Figure 27. As shown in Figure 27, the table consists of items such as individual ID, learning content, REGLINE YINTC, REGLINE GRAD, and REGLINE ESTI.

[0237] The learning content 2111 indicates the content set, REGLINE YINTC indicates the intercept b of the regression line yi, REGLINE GRAD indicates the coefficient "a", and REGLINE ESTI indicates yi. That is, in this embodiment, as shown in FIG. 28, the web-side application server sequentially generates a terminal-side generated one-day presentation list SFi (FIG. 7(b)) for each day (in FIG. 28, the 1st, 2nd, . . . 20th days are shown as examples) based on the local-side presentation list TRi (TR1, TR2, . . . : 60 weeks) prepared on the local server 400 side, and transmits it to the learner terminal 10.

[0238] In this case, for words for which responses are being collected for the first time (words for which implicit memory confirmation has not yet been completed), the time field should be left empty (□: writable). However, depending on the task (D, K, F, etc.), some value may be written in advance to the response (not writable), so no display control code is attached to it, but it is shown as an example empty (□).

[0239] For example, on the first day, the first serial (225) is sent to the learner terminal 10 in the terminal-side generated one-day presentation list SF1 with an empty response (□) and an empty time (□). On the fifth day, it is determined that latent memory has been achieved, and the time is written with ((-2): web-side decision display skip coefficient βi) and sent. From then on, all the time values ​​of serial (225) are set to ((-2): web-side decision display skip coefficient βi) and sent (SF5, ...SF20...).

[0240] That is, from the fifth day onwards, the learner terminal 10 transmits serial (225) with all the display removed as TASK "2" (-2). That is, the latent memory degree confirmation unit (latent memory analysis process) of the web-side application server performs the process shown in FIG. Figure 29 shows the reaction list H4 for the 4th day, the reaction list H10 for the 10th day, and the reaction list H20 for the 20th day from the learner terminal 10, and the terminal-side generated one-day presentation list SFi shows the reaction list H9 for the 9th day and the reaction list H19 for the 19th day.

[0241] In FIG. 29, the serial is explained as "225:Temple". The latent memory confirmation unit of the web-side application server sequentially defines the score Gmi of the response of serial "225:Temple" included in the reaction data hi in the score definition coordinate system HZ1 [serial is "225:Temple": the horizontal axis is the number of times the content is presented in one day Xm] each time a reaction list Hi (H1, ··H10, ··H20, ··) is sent (however, if time is set to (-1) or (-2), it is not defined).

[0242] In addition, Figure 29 shows the reaction list H4 on the fourth day when h1 is sent as [serial(300:about), response(1), time(0.8 seconds)], h2 is sent as [serial(225:temple), response(3), time(1.8 seconds)], h30 is sent as [serial(225:temple), response(3), Time(-1)], and h31 is sent as [serial(225:temple), response(3), time(-1)] (this is an example where on the fourth day the learner terminal 10 received the highest score (3) for all serial(225:temple) from h2 onwards).

[0243] Also, the reaction list H10 on the 10th day is an example where h10 is [serial(300:about), response(2), time(1.2 seconds)] and h11 is [serial(225:temple), response(3), time(-2)], ... Furthermore, the response list H20 on the 20th day is an example where h60 is sent with [serial(111:desk), response(2), time(5 seconds)] and h61 is sent with [serial(225:temple), response(3), time(-2), ··].

[0244] Then, the application server's regression line calculation process averages the responses [(response(1), ..., response(3), ..., response(2))] of [serial(225:temple)] for one day in the coordinate system for defining grades HZi (where the horizontal axis is the number of times the content is presented in one day Xm), and calculates the average individual grade Pm for each day (study session) for this content [serial(225:temple)] by averaging these responses [(response(1), ..., response(3), ..., response(2))].

[0245] Then, this individual's average performance rating Pm is sequentially defined in the coordinate system Vci (specifically, a function) for confirming the degree of latent memory of serial (225:temple) where the horizontal axis is the number of presentation cycles Xi (1st day, 2nd day, ...). Then, each time the individual grade average Pm is defined in the coordinate system Vci for checking the latent memory level (for each learner and content), the regression line y (y=axi+b) is sequentially calculated. Then, each time this regression line y (y=axi+b) is calculated, it is determined whether or not the first condition, which states that the highest dotted line FMa (rating value "3") is reached at the specified number of presentation cycles Xi, is met.

[0246] If it is determined that the first condition is met, it is determined whether the second condition is met, which is that the individual's average performance rating Pm is the highest score ma (rating of 3) consecutively from the specified presentation cycle number Xi (e.g., the 10th day) until the third cycle (e.g., the 8th day). Then, if it is determined that both the first condition (the highest point is reached three times in a row) and the second condition (the regression line reaches the highest point) are met, all subsequent terminal-side generated one-day presentation lists SFi having the same serial (225:temple) are retrieved from the terminal-side generated one-day presentation list SF (SF10) for the specified cycle, and (-2) is written into the time of the terminal-side generated one-day presentation data Sfi having the same serial (225:temple) in these terminal-side generated one-day presentation lists SFi, and the display task code is set to time (see SF9 and SF19 in Figure 29). Note that after writing, the data is set to be unwritable.

[0247] Furthermore, if it is determined that the highest dotted line FMa (rating value "3") has not been reached in the specified cycle Xi (see SF5 in Figure 29), the next reaction data hi (hi = hi + 1) after the specified reaction data hi is specified and the above processing is performed. Furthermore, if either the first or second condition is not met or both are not met in the specified cycle Xi (see SF5 in Figure 29), nothing is done, and the next reaction data hi (hi = hi + 1) after the specified reaction data hi is specified and the above processing is performed.

[0248] For example, the response list H4 on the 4th day is sent as follows: h1 is [serial(300:about), response(1), time(0.8 seconds)], h2 is [serial(225:temple), response(3), time(1.8 seconds)], h30 is [serial(225:temple), response(3), time(-1)], and h31 is [serial(225:temple), response(3), time(-1)]. If this is the case ((this is an example where on the fourth day the learner terminal 10 has given the highest score (3) to all serial(225:temple) from h2 onwards), and it is determined that serial(300:about) and serial(225:temple) do not satisfy the above-mentioned latent memory attainment conditions, the one-day presentation list SFi generated for the terminal side will be sent with the Time for serial(300:about) and serial(225:temple) left blank (□).

[0249] In this example, serial(300:about) in the terminal-side generated one-day presentation list SF5 in Figure 29 is sf10 indicating the 10th, and serial(225:temple) is sf11 indicating the 11th. In other words, even if serial(225:temple) is judged to be the highest score ma(3) on the learner terminal 10 once, twice, or three times discontinuously in cycles before the 5th day, if the latent memory attainment condition is not met in the terminal-side generated one-day presentation list SF5 in the 5th cycle by the above processing, (-2) will not be written.

[0250] On the other hand, for example, in the reaction list H8 (not shown) for the 8th day to the reaction list H10 for the 10th day, if the reaction data hi having the serial (225:temple) of each list is transmitted from the learner terminal 10 in response (3) and it is determined that the above-mentioned latent memory attainment condition is met, assuming that the same serial (225:temple) one-day terminal-side generated presentation data Sfi is also present on the 11th day, the one-day terminal-side generated presentation data Sfi of serial (225:temple) of the 11th one-day terminal-side generated presentation list SF11 is transmitted with Time defined as "-2 (unwritable)". At this time, "-2 (unwritable)" is written to all the terminal-side generated one-day presentation data Sfi in serial (225:temple) of the terminal-side generated one-day presentation list SFi for the subsequent cycles (see SF19).

[0251] Therefore, the reaction list Hi from the 11th day onwards will be sent to the web side with the time of the one-day presentation data Sfi generated for the terminal side of serial (225:temple) set to -2 (see H20).

[0252] [Supplementary explanation of schedule conditions and schedule tables] (Measures for incorporating schedule factors into analysis)

[0253] The schedule of this embodiment uses event cycle units, intervals, condition units, etc. that define the schedules described in Japanese Patent Nos. 3764456, 4391474, 5130272, etc. An overview will be given below. As shown in Figure 30, Schedule A, where you study once a day for 10 days, and Schedule B, where you study five times on the first and sixth days, both involve a total of 10 study sessions, but the quality of the learning in each schedule is clearly different.

[0254] These schedules can be divided into two categories: timing conditions (in Schedule B, when one event occurs, there is a five-day interval between that point and the occurrence of the next event) and presentation conditions (in this example, the dimension representing the number of times of learning (strength of learning)).

[0255] In other words, if the minimum period (presentation unit) during which an event related to one type of English word occurs is one day, the presence or absence of the event during that minimum period included in 10 days is represented (encoded) by corresponding it to 1-0 regardless of the number of repetitions or the types of various conditions, thereby expressing (encoding) the timing conditions for the occurrence of a specific event during that period. Specifically, as shown in Figure 3230, Schedule A is represented as (1, 1, 1, 1, 1, 1, 1, 1, 1), and Schedules B and C are represented as (1, 0, 0, 0, 1, 0, 0, 0, 0).

[0256] According to this, Schedules B and C are the same in terms of timing conditions, but differ in the dimension of the presentation condition, that is, the number of repetitions. There are many possible types of presentation conditions, such as the number of repetitions, the presentation time (for example, presentation for 3 seconds), the processing required at the time of presentation (e.g., assessment of familiarity with English words, individual assessment of learning achievement, etc.), and the characteristics of the order in which content is presented within the presentation unit. There are many possible ways to express these, but by combining the notation of these presentation conditions with the notation of timing conditions, we believe it is possible to express all events that occur at unspecified times (in other words, the content of an individual's experience).

[0257] To summarize, in this example, we first define the minimum period during which an event occurs as a presentation unit, and then propose dividing the conditions conventionally referred to as schedules into two independent dimensions, timing conditions and presentation conditions, and examining their influence. If necessary, it is also possible to consider content attributes (difficulty) as another dimension.

[0258] (Event cycle unit settings) We envision an event cycle unit as a framework for controlling the timing of events related to a huge amount of content. If it is difficult to present a huge amount of content all at once, it will naturally become necessary to scatter the content over a certain period of time. This period will be called an event cycle unit. One point to note is that there are restrictions on how this period can be set.

[0259] In other words, in order to generate a schedule in which events related to all content occur repeatedly at a fixed timing, it is necessary to set an event cycle unit that is a fixed period that is longer than the presentation unit described above and shorter than the shortest interval set within the expected period, and to arrange the content so that a specific event related to each content occurs once within that event cycle unit.

[0260] Therefore, in this case, the timing condition is composed of the elements of the presentation unit, the event cycle unit, and the interval.

[0261] For example, consider the case where you study English vocabulary for one year, studying every two months for the first six months, and once a month for the remaining six months. In this case, the minimum interval is one month, so in order to control the timing conditions, the period for scattering (placing) content must be a maximum of one month. If you set an event cycle unit that exceeds one month, it will be impossible to strictly control the timing of content-related events.

[0262] Generally, all event cycle units are fixed from the point when the event under consideration first occurs, and all event cycle units are of the same length.

[0263] 31 shows two examples of event cycle units that can be set in accordance with schedule B. In this embodiment, each content is arranged so that it is presented once in each of these units.

[0264] As shown in Figure 31, the interval between the presentation (learning or event) of a certain content (English word) and its re-presentation is called an interval, and a period longer than a presentation unit and shorter than the shortest interval is fixed as an event cycle unit mi, the start of which is set to coincide with the start of the event occurrence timing (it does not necessarily have to coincide, but the start of the event cycle unit after the first event cycle unit is specified as the interval after the start of the first event cycle unit). Also, a certain content is presented once within one event cycle unit mi.

[0265] Depending on how you look at it, you could also consider the event cycle unit mi to be the five lessons that take place on the first day of Schedule B, but in that case, the interval (minutes, seconds, etc.) between the first and second lessons would be defined as the smallest unit of one event cycle unit mi = presentation unit (bi) (in other words, it is also necessary to clarify what the smallest unit of an event cycle unit = presentation unit will be). However, when considering a general learning event and its effects, as shown in Figure 30, it is sufficient to assume that the smallest unit of an event cycle unit mi = presentation unit (bi) is about one day.

[0266] FIG. 32 is a diagram showing the uniform arrangement of presentation conditions and the uniformization of condition units and presentation units. The presentation conditions 1 to 16 indicate the degree of difficulty, with 1 being the number of times easy content is presented (1) and 16 being the number of times the most difficult content is presented (16). For example, this means that one easy piece of content and 16 of the most difficult content are presented per day.

[0267] Here, returning to FIG. 31, an explanation will be given again. Schedules B and C in FIG. 31 are when one day is set as the smallest unit (presentation unit) bi of the event cycle unit mi, and represent that the timing conditions Ai are the same but the presentation conditions Bi are different.

[0268] The presentation condition Bi is not limited to the number of times, and various variations such as presentation time and response method can be considered.

[0269] In other words, if the schedule is only concerned with indicating the number of days between events and intervals, and events are represented as 1 and 0, then "5" in Schedule B can simply be represented as "1." Since one piece of content is presented once within one event cycle unit, "1" is set on the first and sixth days, as shown in Figure 30.

[0270] Therefore, the timing conditions Ai of the schedule ABC in Figure 30 and the event presentation conditions can be coded.

[0271] (Content grouping) In order to clarify individual characteristics and general trends, and to make accurate predictions, it is necessary to clarify the quality of the content that requires a response, such as a rating. If the interest is in changes in responses to individual products or English words, there is no need to group them, but in order to see the response trends of an individual, it is desirable to take a representative value of the responses to multiple pieces of content sent to that individual.

[0272] In such cases, multiple similar contents are prepared for each combination of timing condition Ai (pattern of rest intervals in learning, etc.) and presentation condition Bi (number of learning sessions, learning intensity). This is content grouping.

[0273] For example, in the case of English vocabulary, content items are grouped together by certain attributes, such as words of the same level of difficulty, or words required for each grade level (attribute information is added). Note that difficulty here does not only refer to general difficulty, but can also refer to the difficulty level associated with the individual's performance on each piece of content, or the difference between the individual's response to each piece of content and the response of an expert.

[0274] (Content Group: A collection of similar content items) In the case of English vocabulary, it is a collection of content items that can be grouped together by certain attributes, such as words of the same level of difficulty, words required for each grade level, or words required for studying a certain unit.

[0275] As mentioned above, a content item occurs (is presented) only once in one event cycle unit of the timing condition (note that "once" is not a presentation condition).

[0276] (Reasons for grouping content) The reason for grouping content is that the effect of a specific schedule is thought to differ depending on the content. The effect of a schedule on a set of content with multiple attributes is likely to be different if the set of content changes. In other words, the intention is to increase the accuracy of the analysis by separating content factors and schedule factors as much as possible.

[0277] Furthermore, by making the content homogenous and grouping it as finely as possible, it becomes possible to use the obtained data to conversely estimate the relationship between content groups of different quality from the schedule and the reactions to each content group. (Method of determining timing conditions) Although an infinite number of types of timing conditions are possible, in order to make the effect of the event cycle easier to understand visually, it is effective to initially set equal intervals and presentation conditions, such as once every 10 days or once a month.

[0278] In order to provide data that is easy to understand even for early raters, it is advisable to start by collecting data under relatively simple timing conditions such as those described above. In any case, it is advisable to set the timing conditions and presentation conditions that determine the event cycle unit in a well-planned manner.

[0279] To make it easier to visually understand the effect of intervals on timing conditions, it is initially effective to set equal intervals, such as once every 10 days or once a month (see Figure 31). (Uniform arrangement of presentation conditions within each event cycle unit) In a specific event cycle unit, the presentation conditions to be considered there are arranged as uniformly as possible within one event cycle unit, and then content items are allocated there.

[0280] Taking English vocabulary learning as an example, if items with a particular presentation condition are unevenly distributed between the beginning and end of a certain event cycle unit, and a situation arises in the middle of the event cycle unit where learning is no longer possible due to illness or other reasons, the timing conditions will differ slightly depending on the condition. Also, as will be discussed later, when setting an evaluation event in a shorter period than the event cycle unit and estimating the effects of the event by grouping content by presentation condition, it becomes difficult to strictly control the period between the event for which effects are considered and the evaluation event.

[0281] To prevent this, the presentation conditions should be made as uniform as possible within each event cycle unit. More specifically, it is effective to provide the condition unit described below.

[0282] (Method of uniform arrangement of presentation conditions) In addition to randomly arranging presentation conditions, there is another method for uniform placement: Divide the event cycle unit into fixed periods (called condition units), and within each condition unit, all presentation conditions (not content items) or presentation conditions or attribute conditions for which you want to compare the effects of events that occur under certain timing conditions appear once each (see Figure 32).

[0283] In this case, the timing condition is composed of elements such as a presentation unit, a condition unit, an interval, and an event cycle unit.

[0284] Figure 32 shows the conditions under which an event cycle unit of approximately 2 months (48 days) is repeated, with 16 conditions set as presentation conditions Bi, with the number of presentations (learning times) ranging from 1 to 16. In this example, the first event cycle unit mip (48 days) is divided into 4-day intervals, and a total of 12 condition units fi are created.

[0285] For example, on the first day, the learning times are assigned as 1, 8, 9, and 16, and on the second day, the learning times are assigned as 2, 7, 10, and 15, so that all 16 conditions appear once in one condition unit (4 days).

[0286] Similarly, all 16 conditions appear once in the second conditional unit fi on days 5 to 8. Similarly, 12 conditional units fi are repeated in one cycle (48 days).

[0287] The number of content items assigned to one presentation condition Bi of each condition unit fi does not have to be the same. Also, in Figure 32, the order of the condition combinations between days 1 to 4 and days 5 to 8 is different, but this is to reduce the influence of the order in which the presentation conditions Bi appear. The order in which the conditions appear over the four days can be random or a more strict counterbalance method can be used.

[0288] (Fixing the order in which content items are presented) In each event cycle unit, the order in which the content items are presented is fixed, and the content items are arranged so that they are presented in approximately the same order in any event cycle unit.

[0289] If the timing at which a particular content item is presented differs within each event cycle unit, an error will occur in the schedule conditions (schedule conditions are a general term for presentation conditions and timing conditions).

[0290] For example, even if data is collected under timing conditions where an event is repeated at one-month intervals, if a content item presented on the last day of the first event cycle unit is presented on the first day of the second event cycle unit miq, it will be difficult to classify the response to that content item under the timing conditions of a one-month event cycle unit repetition.

[0291] By fixing the order of presentation of content items within an event cycle unit, the effect of any variations in presentation within each unit can be kept relatively small.

[0292] (Homogenization of each presentation unit) Here, a group of content items presented within the smallest unit of an event cycle unit is called a presentation unit Pi. In the case of English vocabulary learning, this is like a group of English words presented on a certain day.

[0293] The combination of presentation conditions Pi, the number of content items, or the combination of schedule conditions is adjusted so that the quality of events required for each presentation unit Pi (total number of presentations and total required time) is as equal as possible.

[0294] Of course, various adjustments can be effective, such as randomizing the presentation order within a presentation unit, or presenting the same content as far apart in time as possible.

[0295] This operation means that the events occurring within this presentation unit Pi do not vary significantly for each presentation unit Pi. For example, if one presentation unit Pi is required to be studied every day, as shown in Figure 32, and one day ten English words with a study frequency condition of 16 times appear, and another day ten words with only one time appear, the study conditions for each content item cannot be said to be equal (naturally, the time required for study will differ). In such cases, when estimating the effect of each event, it is not desirable to simply distinguish between them based on the timing conditions, presentation conditions, and attribute information. To prevent this, as an example, in Figure 32, the total number of learning times (presentation conditions) within a presentation unit is allocated so that they are approximately equal.

[0296] Of course, this adjustment can be made not only by the presentation conditions, but also by the number of content items allocated to each presentation condition or by a combination of different schedule conditions.

[0297] Even if the schedule and presentation conditions are the same, if the response burden required in one presentation unit Pi differs from that required in another presentation unit, the meaning of the response within the presentation unit will differ. Therefore, it is necessary to eliminate this possibility as much as possible, and this is a measure to achieve this.

[0298] (Miniaturization of the presentation unit) In order to collect response data for various combinations of schedule conditions, presentation conditions Bi, content group attributes, etc., it is desirable to keep the time spent responding to one presentation unit as short as possible (if it takes too long, it will not be possible to guarantee consistent presentation conditions and the burden on the person responding will be increased).

[0299] Therefore, it is effective to lengthen the event cycle unit, reduce the number of presentation conditions Bi per presentation unit Pi, or reduce the number of content items allocated to the presentation conditions Bi.

[0300] If one content item is assigned to each of the presentation conditions Bi of the presentation unit Pi on the first day (1, 8, 9, and 16 times in Figure 32), the total number of learning times per day will not be so large. Of course, in general, the event cycle unit mi (mip, miq) and the condition unit fi may be made longer.

[0301] Furthermore, individual evaluation criteria change daily. To account for and eliminate this variation in analysis, a method is used in which content with multiple presentation conditions is prepared within the same presentation unit, and the difference in response to that content is used as an index of the strict condition difference.

[0302] (A method to eliminate the influence of everyday encounters and reveal general trends) The content items can be allocated to the arranged presentation conditions by randomly allocating them to each rater, or by using a counterbalance method as used in conventional experimental research.

[0303] Many of the content items that are presented in a planned manner are encountered by chance in everyday life. Although it is difficult to control this, in order to grasp general trends, it is effective to equalize or randomize the likelihood of encountering a specific content item by chance in everyday life, and then take a representative value from the data of multiple people when analyzing.

[0304] Taking the schedule in Figure 32 as an example, when participants are asked to evaluate their own learning achievement for the words in the one-time learning condition presented on the first day of the third month of the second event cycle unit, the response shows the influence of the learning event that was learned once two months earlier.

[0305] Similarly, if a subject is asked to rate the words in the 16 learning conditions, the effects of the 16 learning events two months earlier will be apparent. If one content item is assigned to each learning session, and a similar rating event continues for four days (one condition unit), response data for one content item will be collected for each of the 16 presentation conditions.

[0306] If this continues for another two months, at the end of the second event cycle unit (the end of the fourth month), responses to 12 content items with the same schedule conditions will be collected for each of the conditions from 1 to 16. By aggregating these responses, it will be possible to obtain a representative value for the responses to content items with the same schedule conditions.

[0307] (test) Tests are generally conducted over a short period of time, and in this case, the method for setting the event cycle unit must deal with the difference in the time interval between the learning event and the evaluation event such as a test (hereafter sometimes referred to as a test). For example, in Figure 32, if learning is conducted in the first and second months and the effect is measured on a day in the third month or later, it would be difficult in terms of time to use all of the content used in the first and second months for the test, so some content would need to be extracted to create the test items. In that case, if the test were to be conducted on the first day of the third month, for example, the interval between learning and the test would vary from 1 to 24 days.

[0308] In this situation, if a test were to be constructed using content learned in the first half of the first month and content learned at the end of the second month and the scores compared equally, the effects of the difference in the interval between learning and the test would be mixed into the scores, making a strict comparison difficult. There are two methods for making an accurate comparison that take into account the effects of the difference in the interval between learning and the test.

[0309] One method is to extend the learning schedule shown in Figure 32 up to the sixth month, and administer tests in the third, fifth, and seventh months to compare results after repeated learning. The test date is the same every month, and the test items are constructed by selecting the same number of pieces of content from the content items at the same time within the event cycle unit. For example, if tests are to be administered on the first day of the third and fifth months, the test items are constructed by selecting one piece of content each from the content items in the presentation units on the sixth, twelfth, eighteenth, and twenty-fourth days of the first month (see Figure 33).

[0310] In this case, if test items for the third and fifth months are extracted in the same way, the tests are administered, and a representative value is taken for all of them, the test scores for the third and fifth months will show results that are similarly affected by the difference in the period from study to test. This is a method similar to the counterbalance method traditionally used in experimental psychology, but applied to timing conditions.

[0311] While this method is effective for examining the effects of learning from month to month, it makes it difficult to compare the effects of presentation conditions. For example, if a test is administered on the first day of the third month, with many words in the 10-time condition at the beginning of the first month and many words in the 1-time condition at the end of the second month, testing the words in the 10-time condition and the 1-time condition on the first day of the third month and comparing the results would not allow for an equal assessment of the effects of both conditions. A method to resolve this issue is to use the scheduling method of setting condition units, and then, when constructing the test, extract the same number of items from the same condition units for the conditions being compared, construct the test, and then compare the responses together.

[0312] According to this, the interval between learning and testing can be considered equivalent under all presentation conditions (see Figure 33). Extracting the same number of test items from the same condition unit means that the differences in the interval between learning and testing are equal, and this has the same effect as the previous method (applying the counterbalance method to timing conditions). More importantly, by taking advantage of the characteristic of the condition unit that all presentation conditions are included at least once, it becomes possible to compare the effects of learning under multiple presentation conditions by offsetting the factors of the interval between learning and testing. In other words, this is also an evaluation method that can only be easily analyzed once the condition unit has been set.

[0313] (Adjusting different schedule conditions) for example, In the example test configuration, if tests are administered at 3, 5, and 7 months to continuously measure the impact of a scheduled event, the items used in the test will be affected by the event at the time the test is taken. Therefore, when observing how learning performance changes over the months, it becomes difficult to detect the pure effect of the learning.

[0314] In such cases, in addition to the content that holds learning events according to the schedule shown in Figure 32 (with the condition of studying one day every two months), there is also separate content that holds test events according to the same schedule (of course, a different schedule is also acceptable), and the two events can be run side by side to allow participants to experience the events, with the results being compiled separately.

[0315] In this way, it is entirely possible to prepare multiple schedules and content for individuals to experience and run them in parallel over the same period. In particular, when individuals experience similar events using the same type of content, it is possible to collect and analyze response data for multiple schedules individually, even if the individuals responding are not aware that they are being asked to respond under different schedules.

[0316] In this case, it is desirable to use different content items and make the individual burden within a presentation unit (one day in the case of Figure 32) the same for every presentation unit as much as possible.

[0317] This method is particularly effective in identifying the relationship between responses to events occurring under multiple schedule conditions and inferring the pattern of response change under one schedule condition from the pattern of response change under another schedule condition.

[0318] (How to configure timing conditions to accommodate different schedules) When making the above predictions, it is desirable to create as many schedule conditions as possible in advance and accumulate data on reactions to events. In this case, an infinite number of timing conditions can be considered, and a timing condition must be described for each one. For example, expressing a large number of timings as "once a week" or "once a month" would not be able to efficiently express such timings. Furthermore, the method of determining the minimum period of a presentation unit and expressing timing based on whether or not an event occurs within that period (expressing it as a pattern of 0s and 1s) becomes more difficult to express as the total period becomes longer.

[0319] Furthermore, when multiple schedule conditions are applied to the same learner simultaneously, it is not efficient to create a schedule table each time that shows the contents of the daily learning conditions. In particular, the following method is used to automatically generate a table of schedule conditions in which the same timing conditions occur repeatedly.

[0320] The presentation unit is the smallest unit, and timing conditions are expressed by the length of the event cycle unit (E), the length of the interval (I), the length of the condition unit (J), and the length of the delay (D).

[0321] As shown in Fig. 34, for example, one week is six days, and one month is four weeks, that is, 24 days. Various schedules are adjusted to these months and days. The A1 schedule is expressed as "E024 I024 J002 D000," which means that if the basic presentation unit is one day, the event cycle unit is one month, the interval is one month, the length of the condition unit is two days, and the length of the delay period from the start of learning (in this case, the first day of the first month) to the start of the first A1 condition is 0 days (no delay).

[0322] The horizontal bars in the graph of the figure indicate that the scattered content is scheduled to be presented (occurrence of an event) within the interval unit.

[0323] Similarly, the B1 schedule indicates that the event cycle unit is 2 months, the interval is 2 months, the length of the condition unit is 4 days, and the length of the delay period from the start date of learning to the start of the first B1 condition is 0 days.

[0324] As an example of a delay period, the C2 schedule is expressed as "E006I006 J001 D006." This schedule indicates that the event cycle unit is 6 days, the interval is 6 days, the length of the condition unit is 1 day, and the length of the delay period from the start date of learning to the start of the first C2 condition is 6 days.

[0325] The above example shows a schedule where events are scheduled to occur every day, but if necessary, it is also possible to wait a short period after completing the first learning before starting the second learning. Examples of such schedules are D1, D2, E1, E2, and E3.

[0326] For example, the schedule for D1 is expressed as "E006 I006 J001 D000E004 I008 J000 D000".

[0327] The first half of this notation, i.e., "E006 I006 J001 D000," will be called the basic unit. This basic unit can be interpreted in the same way as the notation described above, indicating that the event cycle unit (basic event cycle unit) is 6 days, the interval (basic interval) is 6 days, the length of the condition unit (basic condition unit) is 1 day, and the length of the delay period (basic delay period) from the start of learning to the start of the first D1 condition is 0 days.

[0328] On the other hand, the latter half, i.e., "E004 I008 J000 D000," will be called the sub-unit. The meaning of the contents of the sub-unit is that the basic interval period of 6 days in the basic unit is regarded as the presentation unit, which is the basic unit when describing a schedule, and it shows what the schedule would look like if it were re-expressed using that 6-day presentation unit. When the basic interval is regarded as the presentation unit, the interval, event cycle unit, condition unit, and delay period in the sub-units are called sub-interval, sub-event cycle unit, sub-condition unit, and sub-delay period, respectively.

[0329] For example, if the presentation unit on the 6th day is called 1 unit, the secondary event cycle unit is 4 units, the secondary interval is 4 units, the length of the secondary condition unit is 0 units, and the length of the secondary delay period from the start of learning to the start of the first D1 condition is 0 units.

[0330] Next, in the case of a schedule like D2, if we consider that there is a one-month delay period from the start date of learning, and from the second month onwards, events occur for one month, followed by a one-month break, and this cycle is repeated, then if the basic interval of 6 days is one presentation unit, the secondary event cycle unit is 4 units, the secondary interval is 8 units, the length of the secondary condition unit is 0 units, and the length of the secondary delay period is 4 units. In this way, the length of the delay period is expressed in secondary units, and is not expressed in basic units, but is set to 0.

[0331] Therefore, it would be written as "E006I006 J001 D000E004 I008 J000 D004".

[0332] For D2, the basic event cycle unit is 6 days, the basic interval is 6 days, the length of the basic condition unit is 1 day, and the length of the basic delay period from the start of learning to the start of the first D1 condition is 0 days.If the basic interval of 6 days is considered to be 1 presentation unit, the secondary event cycle unit is 6 units, the secondary interval is 6 units, the length of the secondary condition unit is 0 units, and the length of the secondary delay period is 0 units.

[0333] In the E3 schedule, the basic event cycle unit is 6 days, the basic interval is 6 days, the length of the basic condition unit is 1 day, and the length of the basic delay period from the start of learning to the start of the first E3 condition is 0 days.If the basic interval of 6 days is considered to be 1 presentation unit, the secondary event cycle unit is 6 units, the secondary interval is 6 units, the length of the secondary condition unit is 0 units, and the length of the secondary delay period from the start of learning to the start of the first D2 condition is 8 units.

[0334] The method of interpreting the schedule conditions of E1, E2, and E3 is the same as above.

[0335] As described above, by configuring schedule conditions with a base unit and sub-units, it is possible to efficiently express various schedule conditions. Taking this a step further, for example, by further expanding and configuring the sub-units as a first sub-unit, a second sub-unit, etc., it is possible to express schedule conditions in multiple dimensions, such as a base unit, a first sub-unit, a second sub-unit, etc.

[0336] 35 and 36 show more specific examples of the content DB stored in the content database 1. A pair of English and Japanese content (Q and A) is stored corresponding to a serial number (SERIAL), along with content attribute information (type: TYPE), familiarity rating value (F00), familiarity rating standard value (FNORM) for the target learning group, etc.

[0337] Specifically, as shown in Fig. 37, the schedule table is a table that associates schedule type (SCHE TYPE), serial number (NO), month and day (MONTH and DAY), type (TYPE) which is a condition code of attribute information (difficulty, importance, etc.), condition code (N) indicating the number of contents, condition code (EXP) indicating the type of learning method, dimension of schedule condition (MAX DIM), content identification condition code (EXP COND), condition code (SC COND) indicating the unit of event cycle unit, repeat (REPEAT) which is a condition code indicating the number of repetitions, basic schedule condition code (BASIC COND), etc. Also, a specific example of a content table is shown in Fig. 38.

[0338] This example schedule table shows a portion of a table containing schedules for timing conditions A1, B1, C1, D1, E1, etc. shown in Figure 34. A1 is a condition for studying once a month, B1 is a condition for studying once every two months, and C1 is a condition for studying once a week, D1 is a condition for studying once a week for one month, then taking a break for one month, and then repeating this process again, and E1 is a condition for studying once a week for one month, then taking a break for two months, and then repeating this process again, and

[0339] The minimum unit of the event cycle unit (presentation unit) is assumed to be one day.

[0340] The relationship between the timing condition A1 and the content will be explained with reference to FIG. 39 (the content table is, for example, FIG. 38). For the English words in the attribute condition LB1, in the T condition among the presentation conditions, it is assumed that the drill learning number conditions will be carried out from 1 to 8 times, and five contents (English words) will be presented in each drill learning number.

[0341] In addition, in the D condition, it is assumed that the drill learning conditions will be carried out from 1 to 8 times, and that three contents (English words) will be presented in each drill learning.

[0342] Therefore, in the T condition, the number of contents required to be presented is 5, so when the number of drill studies is 1, the total number of drill items is 1 x 5 = 5, and when the number of drill studies is 2, the total number of drill items is 2 x 5 = 10. Also, in the D condition, the number of contents required to be presented is 3, so when the number of drill studies is 8, the total number of drill items is 3 x 8 = 24.

[0343] The results of this calculation are shown in the column labeled "Total number of drill items," and the total is displayed on the far right.The number of content items required for one condition unit (2 days) is 64, as can be seen from the diagram, so the number of items required for one event cycle unit (24 days) is 64 x 12 = 768 items.

[0344] In addition, in order to arrange the conditions so that the total number of presentations of content items within a presentation unit is equal (so that the learning and test conditions are as equal as possible), the learning count conditions are assigned so that the total number of learning times on the first day and the total number of learning times on the second day are both 18.

[0345] For example, in the schedule table in Figure 37, the event cycle unit periods are a mixture of one month, two months, and one week, so a field called SC COND is provided to indicate which day of which event cycle unit the event occurs in that schedule condition. For example, if "SC COND" for C1 is expressed as "0203," it would be the third day of the second event cycle unit, or the ninth day counting from the start date of study.

[0346] The type of learning method (EXP) mentioned above is one example of a presentation condition, and there are, for example, four types (TDBF). Type 1 (T) is a learning method in which, for example, when learning paired-associate English translations of Japanese, English words are shown before the learning and the familiarity is assessed (recognition test). Type 2 (D) is a learning method that requires recognition tests and drill learning. Type 3 (B) is a learning method that involves drill learning without any recognition tests. Type 4 (F) is a learning method in which, for example, predetermined content is selected from content that has been assessed for familiarity at the end of each month and tested. The presentation method is also different from TDB.

[0347] The number of repetitions (REPEAT) is also an example of a presentation condition, indicating the number of times a presentation unit is repeated. Other presentation conditions include the presentation time, the way in which learning items are emphasized, and the presentation order, such as determining the order in which the same content is presented as far apart as possible in the presentation list. These learning methods, number of repetitions, etc., correspond to presentation conditions. In other words, by encoding various presentation conditions, detailed predictions for each condition become possible.

[0348] The condition code (TYPE) in the attribute information contains symbols such as W1 and S1. In the case of learning Japanese translations, the first letter W indicates the method of presenting an English word and learning its Japanese translation, while S indicates the method of presenting a sentence and learning its Japanese translation, and the second number indicates the number of different Japanese translations. For example, W3 means that there are three different Japanese translations for the English word.

[0349] The content identification condition code (EXP COND) in the schedule table is a collection of codes for various conditions associated with serial numbers. For example, "0101A=W1T=DR=1ST=ADSC=0101J=01" for serial number "1" indicates an event held on the first day of the first month, with a word attribute of type W1, learning method D (also called presentation condition 1), one repetition count (R=1; also called presentation condition 2), and the first day of the first event cycle unit (SC=0101), and indicates that this is schedule data that provides one piece of content that meets the conditions of the first condition unit.

[0350] FIG. 40 shows the learner's personal data obtained as a result of the tests actually conducted using the above processes, and their prediction functions (a simple regression line drawn for each user word).

[0351] Figure 40(a) shows the change in grades when learner X continues to study word A for six weeks under schedule condition C with a one-week (six-day) interval, and then studies word B. The regression line (yai, ybi) shows the grades over one month, for example. Figure 40(b) shows the results when studying under schedule D. As these figures show, grades are steadily improving. In other words, it is possible to predict learning achievement. Of course, conversely, it is also possible to predict changes in performance in the short interval condition from the long interval condition. [Explanation of symbols]

[0352] 10. End of Learner 100 web servers 200 App Server 300 Web side DB server 400 Local Server 500 Local DB Server

Claims

1. A method for analyzing implicit memory, comprising the steps of: connecting a learner's terminal with a server at an analysis center via a communication network; transmitting a presentation list and a content list for each learning session, which are generated by the server based on schedule conditions for learning learning content at regular intervals over a learning period according to the learner's attribute information, to the learner's terminal; and analyzing the learner's implicit memory achievement level for the learning content based on the learner's response data to the learning content contained in the response list (transmission instruction) from the learner's terminal, It is a method of describing experiences (a set of consecutive events) that precisely describes the expression of "who," "when," "what," and "how" an action was taken, and the expressions of "when" and "how" are expressed as timing conditions based on whether or not a specific event occurred within a presentation unit, and the number of times the event occurred within the presentation unit is separated into a presentation condition called the number of repetitions, thereby simplifying the expression of "when" and at the same time making the expressions of "when" and "how" independent. The presentation unit is a minimum period of the schedule condition, and the presentation list is associated with this presentation unit to simplify the expression of "when," and further, the presentation condition is defined in the presentation list to separate the expressions of "when" and "how," The analysis center transmits to the learner terminal a task process that executes a process that specifies which content elements to present, in what order, and how to record responses for a set of various content elements, and then includes the task code in the presentation conditions and transmits the same to the learner terminal; Display the content on the learner's terminal and acquire the response, and if the highest score is obtained, assign a skip code to skip displaying and recording the subsequent same content in the presentation list; The analysis center receives the response list for the learning session, and if the response is the highest point for a predetermined number of consecutive sessions compared with past response lists and the regression line for this learning session reaches the highest point, assigns a skip code to the same learning content in the presentation list for each learning session over the subsequent learning period; This is the presentation list, which is sent to the learner's terminal for each learning session.

2. A method for analyzing implicit memory, comprising the steps of: connecting a learner's terminal and a server at an analysis center via a communication network; transmitting to the learner's terminal a presentation list for each learning session and a content list by difficulty level, the presentation list being generated based on a schedule condition for learning learning content corresponding to the learner's attribute information at regular intervals over a learning period; and analyzing the learner's implicit memory achievement level for the learning content based on the learner's response data to the learning content contained in the response list from the learner's terminal, The presentation list includes: The presented data, which is a set of a serial code of the learning content in the content list, a column for acquiring an individual evaluation score, a column for acquiring a reaction time, and a task code of the TASK process, is arranged for the same learning content and different learning content, The learner terminal includes: (A1) The TASK process is composed of multiple tasks that execute how to display the learning content and how to acquire data according to a set task code without separating them; (A2) A skip task process for skipping the designated presentation data and designating the next presentation data; (A3) A process of reading the reaction time acquisition column with priority, and if this column is "empty", setting the task code to TASK processing and executing it, and if a negative value is written, setting a skip code to execute the skip task processing; (A4) A process of using a set of the serial code of the presentation data in the presentation list, the individual performance evaluation value in the column for obtaining the individual performance evaluation value, and the reaction time in the column for obtaining the reaction time as reaction data, and transmitting this reaction data as a reaction list to the analysis center in response to a transmission instruction; is downloaded from the analysis center's server, The negative value is When the skip task processing on the learner terminal writes the highest score in the column for obtaining the individual evaluation score, a negative value is written in the column for obtaining the reaction time of all subsequent presentation data having the same serial code in the presentation list, and further, The server of the analysis center Each time a reaction list for each learning session of the learner is received, the reaction list up to a predetermined number of sessions is designated, and if the individual score value in the column for obtaining the individual score value of the reaction data having the same serial code in these reaction lists continuously indicates the highest score, and the regression line of the learning content of this serial code reaches the highest point, it is determined that sufficient latent memory has been reached, and a negative value is written in the column for obtaining the reaction time of the presentation data having the same serial code in all subsequent presentation lists, and these are transmitted to the learner's terminal as a presentation list for each learning session. A method for analyzing implicit memory, comprising:

3. The negative value is The learner terminal uses a first coefficient indicating that it is on the learner terminal side, and the server of the analysis center uses a second coefficient indicating that it is on the analysis center side, which is different from the first coefficient.

3. The method for analyzing implicit memory according to claim 2.

4. The analysis center If the regression line has not reached the highest point, the writing of the negative value is stopped and the next data to be presented is designated.

3. The method for analyzing implicit memory according to claim 2.

5. The analysis center Calculating the continuation of the highest points and the regression line using the reaction data included in the reaction list for the current learning session received from the learner terminal, excluding reaction data containing negative values.

3. The method for analyzing implicit memory according to claim 2.

6. The processing of the analysis center is (B1) A step of storing a reaction list from the learner terminal each time it is received, and specifying past reaction lists up to past learning sessions, including the current reaction list; (B2) The learning content in the current evaluation response data and all past evaluation response data for the same learning content in all subsequent presentation lists are specified, and the average of the individual performance ratings included therein is calculated as the rating for the learning content for the current learning session. (B3) A regression line showing the trend between the grades of the current learning session and each of the grades of the previous learning sessions is calculated. (B4) Each time a rating value for the learning content in the current learning session is calculated, if the rating values ​​for the previous predetermined sessions, including the current rating value, are the highest rating value in succession, and the value of the regression function in the current learning session has reached the highest rating value, it is determined that sufficient latent memory has been achieved, and negative values ​​are written in the reaction time acquisition boxes of all subsequent presentation data that are identical to the learning content in the presentation list. (B5) If the value of the regression line has not reached the highest evaluation value, the writing of the negative value is stopped and the next presentation data is designated.

3. The method for analyzing implicit memory according to claim 2.

Images