Knowledge information providing system

Abstract

To provide a knowledge information providing system that enables a user to intellectually activate his / her leisure by using digital techniques.SOLUTION: A knowledge information providing system 3 includes a server 2 and at least one user terminal 1 connected to the server via a communication network, and provides digitized knowledge information from the server to the user terminal. The server 2 transmits knowledge information selected for each user according to predetermined conditions to the user terminal, acquires the user's reaction to the knowledge information from the user terminal, and provides next knowledge information to the user terminal according to the acquired user's reaction and the predetermined conditions.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a knowledge information providing system. [Background technology]

[0002] For example, the technology described in Patent Document 1 provides an advanced answering machine that plays a timely and customized voice message to a caller according to the presence status of the called party when the caller attempts to contact the called party and fails.Patent Document 2 proposes a technology that receives speech (input) from a user using an online shopping site, provides a response to the speech to a user terminal, converses with the user, and presents products recommended to the user through the conversation. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2007-97162 A [Patent Document 2] JP 2020-57350 A Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, the aging of the population has progressed, and there is a demand for digital technology to be used to intellectually activate the elderly in their leisure time, but there is no system that meets this demand.

[0005] The present invention has been made in consideration of the above-mentioned problems, and its object is to provide a knowledge information provision system that enables users to intellectually activate their leisure time using digital technology. [Means for solving the problem]

[0006] In order to solve the above problems, a knowledge information provision system according to one aspect of the present invention comprises a server and at least one user terminal connected to the server via a communications network, and provides digitized knowledge information from the server to the user terminal, wherein the server transmits knowledge information selected for each user in accordance with specified conditions to the user terminal, acquires the user's reaction to the knowledge information from the user terminal, and provides the next knowledge information to the user terminal in accordance with the acquired user's reaction and the specified conditions. Effect of the Invention

[0007] According to the present invention, knowledge information selected for each user is provided to a user terminal, and the next piece of knowledge information can be provided in accordance with the user's reaction and a predetermined condition. [Brief description of the drawings]

[0008] [Figure 1] 1 is an overall schematic diagram of a knowledge information providing system. [Diagram 2] FIG. 2 is a hardware configuration diagram of the knowledge information providing system. [Diagram 3] FIG. 2 is a functional configuration diagram of the knowledge information providing system. [Figure 4] 2 is an example of a personal attribute database. [Diagram 5] 1 is an example of a quiz database. [Figure 6] 1 is an example of a user reaction database. [Figure 7] FIG. 13 is an explanatory diagram showing the correlation between quiz difficulty and preferences. [Figure 8] FIG. 13 is an explanatory diagram showing the correlation between quiz difficulty and preferences for each user attribute. [Figure 9] 1 is an example of a Venn diagram of user personal attributes. [Figure 10] This is an example of a Venn diagram of the golf population in Japan. [Figure 11] This is an example of non-hierarchical cluster analysis (x-means) by user attributes. [Figure 12] 13 is an example of a quiz screen. [Figure 13] 13 is a flowchart showing an overall process. [Figure 14] 14 is a flowchart showing the assessment process in FIG. 13. [Figure 15] 15 is a flowchart showing the individual assessment process in FIG. 14. [Figure 16] 15 is a flowchart showing a quiz assessment process in FIG. 14. [Figure 17] 13 is a flowchart showing a process for updating a user reaction database. [Figure 18] 13 is a flowchart showing a question setting process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The knowledge information providing system according to this embodiment provides knowledge information to a user, and activates the user's leisure time. Activating leisure time means, for example, satisfying the user's intellectual curiosity and enriching the user's leisure time. In this embodiment, a knowledge information providing system is provided that allows even so-called digitally disadvantaged people, such as the elderly, to easily enjoy quizzes without getting bored, thereby contributing to improving their education.

[0010] The knowledge information providing system of this embodiment analyzes the hobbies and tastes of each user and provides the user with quizzes that the user can enjoy continuously and with interest, thereby enabling the user to improve his / her knowledge without getting bored. Furthermore, the knowledge information providing system of this embodiment monitors the user's accuracy rate and adjusts the difficulty level of the questions to be posed.

[0011] The knowledge information providing system of this embodiment presents a quiz on general knowledge to the user by a voice chat application using, for example, a so-called smart speaker, etc. Therefore, the user can enjoy the quiz on general knowledge without operating a personal computer, and can spend intellectual time.

[0012] The knowledge information providing system of this embodiment performs an assessment to investigate the attributes of the user and to confirm compatibility with the quiz. Furthermore, the knowledge information providing system detects a subjective index by directly asking the user whether he / she likes or dislikes the quiz while the user is answering the quiz, and physiological indexes obtained from various sensors, such as the user's facial expression, heart rate, and brain waves. This allows the knowledge information providing system to continue to provide questions that will continue to interest the user, and to intellectually stimulate the user's leisure time.

[0013] The knowledge information provision system converts the quizzes prepared according to the user's attributes, various biological reactions to the images used in the quizzes, and the results of actions such as users commenting, expressing "likes," or ignoring (ignoring) the quizzes into feature quantities. The knowledge information provision system forms close user network groups by predicting these feature quantities using machine learning methods such as linear regression, logistic regression, boosting tree, and weighted decision tree. The knowledge information provision system analyzes the tendencies of the user network group and presents quizzes that include images and keywords that match the user's hobbies or preferences, customized for each individual user.

[0014] The knowledge information provision system of this embodiment has, for example, the following configuration: The knowledge information provision system includes a terminal with a display (hereinafter, referred to as a smart speaker terminal here) as a user terminal connected to a server having a voice recognition function, and a server capable of communicating with the smart speaker terminal, the server having a voice recognition function and a quiz generation mechanism.

[0015] The smart speaker terminal has a camera that captures an image of the user's face, a microphone (hereinafter, referred to as a mic) that captures the user's voice, and sensors that capture various biological reactions of the user.

[0016] Sensors that obtain the user's biological reactions include, for example, an EEG headset (brainwave detection belt) for analyzing the user's psychology while taking a quiz, a multi-channel near-infrared cerebral blood flow measurement device (optical topography (NIRS) device), eye camera goggles (or glasses, or head-mounted display) for measuring eye movement (movement of gaze), i.e., eye tracking and pupil size and IPD (interpupillary distance), a smart watch, etc. Smart watches measure some or all of the following: electrocardiogram (ECG), electromyogram (EMG), pulse, body temperature (skin temperature), skin microcurrent or skin electrodermal response measurement (degree of sweating), blood oxygen concentration, maximum oxygen intake (VO2 Max), blood pressure, respiratory rate, etc.

[0017] The server has a detection unit that measures the user's state from the smart speaker terminal, a user information database that associates the user's state with the user's reaction to a quiz and accumulates it, a quiz database that stores quiz information for each genre, and a user reaction analysis unit that analyzes the information stored in the user information database using artificial intelligence technology. The server has a voice input / output unit as a data transmission / reception unit that selects the next quiz from the quiz database according to the analysis result, converts the selected quiz into voice using voice synthesis technology and asks it to the user, and detects the user's answer using voice recognition technology. The user reaction analysis unit can include, for example, some or all of an electroencephalogram analysis unit, a cerebral blood flow analysis unit, an eye tracking analysis unit, a facial expression analysis unit, a voice emotion analysis unit, and various vital analysis units.

[0018] The knowledge information provision system randomly presents quizzes to users, obtains the user's physiological indicators (such as brain waves and pulse rate), behavioral indicators (scores, reaction times, etc.), and subjective indicators (reactions such as "likes") in response to the presented quizzes, and stores them in a user information database. The knowledge information provision system uses artificial intelligence technology to analyze the user information and analyzes the genres or trends of quizzes that the user is interested in or concerned with. The knowledge information provision system then selects a quiz that is close to the user's hobbies and preferences from the quiz database and presents it via the smart speaker terminal.

[0019] According to the knowledge information providing system of this embodiment, it is possible to continuously provide quizzes that are likely to interest the user, thereby enabling the user to intellectually spend his / her leisure time. EXAMPLES

[0020] Example 1 will be described with reference to Figs. 1 to 18. As described later, the knowledge information providing system 3 asks quizzes that are in a genre that is popular with the user or a group of users with similar hobbies and interests to the user, and that are neither too difficult nor too easy, so that the user does not get bored and continues answering the quizzes. For example, the question setting control unit 21 asks questions that the user has answered incorrectly in the past at an appropriate frequency repeatedly until the user answers them correctly, and also asks questions that the user has answered correctly in the past at a low frequency, thereby promoting memory consolidation.

[0021] 1 is an overall view of a knowledge information providing system 3. The knowledge information providing system 3 presents, for example, general knowledge quizzes to a user using voice and images, acquires the user's reactions, and reflects them in the next question.

[0022] The knowledge information providing system 3 includes, for example, at least one user terminal 1 and a server 2, and the user terminal 1 and the server 2 are connected to each other via a communication network CN so as to be capable of two-way communication. The communication network CN may be wired or wireless. The communication network CN may be, for example, the Internet.

[0023] Details of the user terminal 1 will be described later with reference to Fig. 2. First, a functional configuration of the server 2 will be described. The server 2 includes, for example, a question setting control unit 21, a data transmission / reception unit 22, a user attribute analysis unit 23, a user response analysis unit 24, a quiz database 25, a user information database 26, a user interaction management unit 27, and a virtual user management unit 28.

[0024] The question setting control unit 21 has a function of selecting a subject quiz from the quiz database 25 based on the user attributes and the user's reaction to the quiz. The data transmission / reception unit 22 has a function of transmitting and receiving data between the user terminal 1 and the server 2. For example, audio data, moving image data, still image data, graphics data, data measured by various sensors, and the like are communicated between the user terminal 1 and the server 2.

[0025] The user attribute analysis unit 23 analyzes user attributes based on some or all of the user attributes stored in the user information database 26. For example, the user attribute analysis unit 23 analyzes to which user attribute cluster the user belongs, and what hobbies or preferences the user has. The analysis results by the user attribute analysis unit 23 can be stored in the user information database 26.

[0026] The user reaction analysis unit 24 analyzes the user reaction based on some or all of the user reactions stored in the user information database 26. For example, the user's emotions, specific fatigue, likes and dislikes, etc. can be analyzed by the user reaction analysis unit 24. The analysis results by the user reaction analysis unit 24 can be stored in the user information database 26.

[0027] The quiz database 25 stores and manages quizzes to be given to users. The quiz database 25 stores cultural quizzes such as art, music, literature, history, geography, climate, natural science, languages, comics, animation, trivia, gourmet food, and games. The quiz database 25 may store not only cultural quizzes, but also jokes, calculation problems, puzzles, and the like.

[0028] The user information database 26 stores and manages information about users. The information about users includes, for example, user attributes and user reactions. The user attributes include, for example, name, sex, age, occupation, work history, educational background, hobbies, preferences, star sign, blood type, Chinese zodiac sign, height, weight, qualifications held, residential area, birthplace, family composition, and other information. The user reactions include the user's biological reactions to the quiz (sweating, body temperature, blood pressure, brain waves, etc.) and the user's behavior in response to the quiz (type and frequency of button operations, etc.).

[0029] The user interaction management unit 27 manages interactions between multiple users U1 and U2 who operate the user terminal 1. The user interaction management unit 27 provides communication means such as electronic bulletin boards, text chat, voice chat, video calls, and e-mails, and promotes interaction between users through educational quizzes.

[0030] The virtual user management unit 28 sets a virtual user for each user attribute cluster, for example. The virtual user management unit 28 manages the virtual users as rivals or virtual friends in a general knowledge quiz.

[0031] 2 is a hardware configuration diagram of the knowledge information providing system 3. The user terminal 1 includes, for example, a processor 101, a main memory device 102, an auxiliary memory device 103, a communication device 104, an input device 105, and an output device 106.

[0032] The input device 105 may include all or some of devices such as a keyboard 1051, a microphone 1052, an electroencephalogram detector 1053, a heart rate electrocardiogram detector 1054, a touch panel 1055, a camera 1056, an HMD (head mounted display: equipped with a camera for eye tracking) 1057, and an optical topography 1058.

[0033] The processor 101, the main memory device 102, the auxiliary memory device 103, the communication device 104, the input device 105, and the output device 106 are connected to each other via a bus 107 so as to be able to communicate with each other.

[0034] The arithmetic processing device 101 is composed of, for example, a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), etc. The arithmetic processing device 101 reads out and executes various computer programs stored in the main storage device 102 or the auxiliary storage device 103, thereby realizing each function provided by the knowledge information provision system 3.

[0035] The main memory device 102 is a device that stores computer programs and data, and is, for example, a Random Access Memory (RAM), a Read Only Memory (ROM), or a non-volatile semiconductor memory (Non-Volatile RAM (NVRAM)).

[0036] The auxiliary storage device 103 is, for example, a reading and writing device for recording media such as a Hard Disc Drive (HDD), a Solid State Drive (SSD), a storage system, an Integrated Circuit (IC) card, a Secure Digital (SD) memory card, or an optical recording medium (Compact Disc (CD), Digital Versatile Disc (DVD), etc.), or a memory area of ​​a cloud server.

[0037] The cache and register of the arithmetic processing unit 101, the main storage device 102, and the auxiliary storage device 103 may be collectively referred to as memory. Necessary computer programs and data are loaded into the memory and executed, thereby executing information processing in the knowledge information provision system 3. Furthermore, computer programs, databases, data sets, etc. may be stored in the memory.

[0038] The communication device 104 is a wired or wireless communication interface that realizes communication with other devices via communication means such as a Local Area Network (LAN) or the Internet, and is, for example, a Network Interface Card (NIC), a wireless communication module, a Universal Serial Interface (USB) module, or a serial communication module.

[0039] The input device 105 is a device that receives input from a user or a device that measures the state of the user using a sensor or the like.

[0040] The output device 106 may include, for example, a monitor display 1061 and a speaker 1062. The monitor display 1061 is, for example, a screen display device (such as a Liquid Crystal Display (LCD) or a Head Mounted Display (HMD)). The output device 106 may also include a printer. The speaker 1062 is a device that provides the user with various information such as the progress and results of processing in the form of voice, music, or sound.

[0041] For example, if the computing power of the user terminal 1 is insufficient, part of the processing can be delegated to the server 2. For example, voice data acquired from the user can be transmitted from the user terminal 1 to the server 2, and the server 2 can perform voice recognition processing.

[0042] The server 2 includes a processor 201, a main memory 202, an auxiliary memory 203, and a user interface device 205 in addition to a communication device 204, and these devices 201 to 205 are communicatively connected to each other via a bus 206. In the drawing, the user interface may be abbreviated as "UI."

[0043] 3 is a functional configuration diagram of the knowledge information provision system 3. The user terminal 1 has functions such as an electroencephalogram detection unit 111, a facial expression detection unit 112, a voice emotion detection unit 113, a cerebral blood flow detection unit 114, an eye movement detection unit 115, and a vital detection unit 116. Data detected by these detection units 111 to 116 is sent to the server 2 via the communication network CN.

[0044] The server 2 has functions such as a question setting control unit 21, a data transmission / reception unit 22, a user attribute analysis unit 23, a user response analysis unit 24, a user information database 26, and a quiz database 25. The user information database 26 includes a user response database 262 and a personal attribute database 261. The user response analysis unit 24 has a function of analyzing the user's response, and includes, for example, an electroencephalogram analysis unit 241, a cerebral blood flow analysis unit 242, an eye tracking analysis unit 243, a facial expression analysis unit 244, a voice emotion analysis unit 245, and a vital analysis unit 246.

[0045] Here, the overall flow of the knowledge information providing system 3 will be described with reference to Fig. 1. The knowledge information providing system 3 is usually used by multiple users U1 and U2. Although Fig. 1 shows two users, in reality, more users use the knowledge information providing system 3.

[0046] The question setting control unit 21 of the server 2 transmits a quiz to the user terminal 1 at a predetermined timing, such as when the user terminal 1 is started up, when instructed by the user, or after a predetermined time has passed, and causes the user to ask the quiz. The user answers the quiz set by the user terminal 1. The user's answer and the user's reaction at the time of answering are transmitted from the user terminal 1 to the server 2.

[0047] The question setting control unit 21 selects and sets quizzes of a suitable level of difficulty that is neither too difficult nor too easy for the user or a group of users with attributes similar to the user from among genres that the user is interested in, or from among genres that a group of users with attributes similar to the user is interested in. Since questions that are all too easy are boring, and questions that are all too difficult discourage the user, a suitable level of difficulty (for example, a national average correct answer rate of 50-60%) is required.

[0048] The question setting control unit 21 re-asks questions that the user has answered incorrectly in the past relatively frequently (e.g., once every 20-30 questions) to encourage review. Conversely, questions that the user has answered correctly in the past are re-asked very infrequently (e.g., once every 100-200 questions) to check whether the question has been retained in the user's memory. If the user answers correctly two times in a row, even if the questions are asked with a considerable gap between questions, it is considered that the question has been retained in the user's memory, and the question setting control unit 21 will not ask the user the question in the future.

[0049] The question control unit 21 very occasionally (for example, once every 50-100 questions) asks questions from genres other than those that interest the user or a group of users with attributes similar to the user, and challenges the user to discover new fields of interest. This makes the quiz system not boring by preventing it from being biased toward a specific genre.

[0050] When the number of questions in the same genre given to the user reaches a certain number (e.g., 10-20 questions), the question setting control unit 21 periodically recalculates the correct answer rate for that genre, and when the correct answer rate exceeds a certain value (e.g., 70%), it raises the level of the quiz to be set. Conversely, when the user's correct answer rate falls below a certain value (e.g., 30%), it lowers the level of the quiz to be set.

[0051] When the number of questions given to the user reaches a certain number (e.g., 10-20 questions), the question setting control unit 21 periodically recalculates the preference index (described below), and when the preference index falls below a certain value (e.g., 50%), reviews the genre of the quiz to be set.

[0052] The question setting control unit 21 stores the results of the quiz set to the user and various reactions of the user in a user sheet of the user reaction database 262 (the column for writing the quiz number is omitted in FIG. 6).

[0053] Various reactions shown by a user when answering a quiz include, for example, subjective indicators, behavioral indicators, and physiological indicators. A subjective indicator is a reaction resulting from a user's subjective judgment of the merits or demerits of a quiz. The user's judgment result can be detected by operation of buttons (such as a Like button, a Boring button, or a Neither button) displayed on the user terminal 2. If the user thinks the quiz is good, the "Like" button is pressed. If the user thinks the quiz is boring, the "Boring button" is operated. If the user does not think the quiz is good or bad, that is, if the user thinks the quiz is not interesting, the user may operate the "Not Interested button." A user may lose interest because the quiz is too difficult, or too easy. A "Too Difficult button" may be provided to indicate that the quiz is too difficult, or a "Too Easy button" to indicate that the quiz is too easy.

[0054] The behavioral index is a user's reaction to a quiz that is not a subjective index. Examples of the behavioral index include the correct answer rate, the viewing rate of explanations and hints, the sound pressure when the answer is spoken, the amount of speech when answering, and the participation rate in the question and answer corner related to the quiz.

[0055] The physiological indices are information obtained from the user's vital signs detected by the user terminal 1. The physiological indices include, for example, various brain waves, heart rate, blood pressure, body temperature, respiratory rate, eye tracking information, and the contraction and contraction state of facial muscles.

[0056] The question setting control unit 21 periodically reads out the user's quiz preference information from the user response database 262 when a certain number of quizzes (e.g., 40-50 questions) have been set to the user, or when a certain amount of time (e.g., one hour) has passed, and analyzes the information using artificial intelligence techniques such as machine learning. There are, for example, two analysis methods. The first is an analysis for grasping the user's own quiz preference tendencies and the optimal difficulty level of quizzes for the user. The second is an analysis for creating a simple quiz preference correlation diagram from information in the user response database 262 of a group of users close to the user. By periodically performing this, the latest analysis data is always kept.

[0057] An example of a user's personal attribute database 261 will be described with reference to Fig. 4. The personal attribute database 261 manages items such as a user ID 2610, height and weight 2611, date of birth 2612, age 2613, blood type 2614, hobbies 2615, academic background 2616, birthplace 2617, family structure 2618, and occupation and work history 2619. In addition to the items shown in the figure, the personal attribute database 261 may also manage, for example, the zodiac sign, exercise habits, qualifications held, and annual income.

[0058] For example, for items such as height, where the values ​​for each item are significantly dispersed if they are considered to differ in units of 1 mm, the number of samples for each item value is appropriately dispersed, for example, by managing them in a range of 10 cm increments. If the item names are diverse and long strings of characters, such as hobbies, qualifications, and place of origin, they may be indexed and stored. Although the date of birth may be managed to the exact date, the attributes of an individual are not represented by the exact date, but by attributes calculated from the birthday, such as the generation (age group), the month of birth, the Chinese zodiac sign, and the star sign. Some of these attributes will not change unless the birthday changes, while others, like age, are updated every year. On the other hand, some, like hobbies and qualifications, change over time. Therefore, individual attributes are divided into static attributes and dynamic attributes, and the database 261 is periodically updated for dynamic attributes that may change, so that the latest data is always stored.

[0059] An example of the quiz database 25 will be described with reference to Fig. 5. The quiz is basically in a one question, one answer format. The answer to the quiz is basically to select one answer from a plurality of answer options.

[0060] The quiz database 25 manages, for example, a quiz ID (not shown), genre 2510, subcategory 2511, number of questions 2512, number of correct answers 2513, national correct answer rate 2514, questions 2515, answers 2516, trivia 2517, and storage locations (not shown) of image files or audio files related to the quiz.

[0061] If the genre or subcategory of the quiz is unknown, it may be automatically classified by a quiz categorization mechanism described below.

[0062] Each time a quiz question is asked, the number of questions asked is counted (this counts questions asked to all users nationwide, not just specific users). In addition, the number of times a question is answered correctly is counted, and the national correct answer rate is calculated and stored. The quiz, the image file to be displayed when displayed on the user terminal 1, the answer options, and the correct answer are stored. Also, for quizzes with explanations, an explanatory text can be stored, and for quizzes with trivia related to the quiz (knowledge about trivial things, general knowledge), multiple pieces of trivia can be stored.

[0063] An example of the user reaction database 262 will be described with reference to Fig. 6. The user reaction database 262 is a database that manages user reactions. The user reaction database 262 manages, for example, a user ID 2620, a category 2621, an index 2622, and a reaction result 2623 for each question.

[0064] A user response category is shown in category 2621. The user response categories include, for example, three categories: a subjective index, a behavioral index, and a physiological index.

[0065] A subjective index is an index that is measured by a user who has made a conscious decision to directly communicate their intention by operating a button or the like. For example, if a user likes a quiz, they press the "Like" button. If the user does not like the quiz, they press the "Boring" button. If neither, the user presses the "Next" button. The "Next" button can also include a reason for skipping the quiz and moving on to the next one. Reasons for skipping can include "not interested," "too difficult," or conversely, "too easy."

[0066] The behavioral index is not the user's explicit will, but is an index based on the user's behavior after being given a quiz. That is, the behavioral index includes, for example, whether the user's answer was correct, the accuracy rate when the question was asked multiple times, the time it took to answer, the percentage of times the explanation button was pressed after answering, the percentage of times trivia was checked, the amount of speech when the answer was given by voice (or when something was spoken while answering), and the sound pressure of the speech (whether the sound pressure was normal or excited, etc.).

[0067] Physiological indices are indices of the user's biological reactions. The user's biological reactions (such as vital signs) can be detected by various signals sent from devices such as a headset attached to the user as part of the user terminal 1, a biosensor such as a smart watch, and a camera that captures changes in the user's facial expression. Of course, if it is not possible to prepare a device that can measure all signals, physiological indices are picked up to the extent possible.

[0068] Possible physiological indicators include brain waves (divided into alpha waves, beta waves, gamma waves, theta waves, delta waves, etc. by frequency), optical topography (NIRS) cerebral blood flow deviation, heart rate, ECG (Electrocardiogram) - RRI (RR Interval: heart rate interval), skin microelectroreactivity (sweating), facial muscle electromyogram (EMG (ElectroMyoGraphy)), blood oxygen concentration (saturation) (SpO2: saturation of percutaneous oxygen), maximum oxygen intake (VO2-Max), blood pressure, respiratory rate, eye tracking (gaze movement), pupil size, IPD (Inter Pupilary Distance), facial expressions or changes in these (deviation from normal) etc.

[0069] These subjective, behavioral and physiological indices are learned by machine learning for each user and each quiz. Basically, the subjective indices serve as training data, and the behavioral and physiological indices of the user when responding with "Like" are observed to analyze the clear difference from when responding with "Boring". Human subjective opinions and behaviors can sometimes result in reactions that differ from what the user really feels. By analyzing the physiological indices, it is possible that even if the subjective and behavioral indices show the same reaction as when pressing "Boring", the user actually feels that they "Like". To correct this, the user's preference (or interest or curiosity) for the quiz is indexed using, for example, the following formula:

[0070] The preference index is defined as follows:

[0071] Preference index = "Like" click rate - "Boring" click rate +Σ(Correlation between the change in various behavioral indicators and the average change in the indicators when pressing "Like") +Σ(Correlation between the change in various physiological indices and the average change in those indices when pressing "Like") -Σ (Correlation between the displacement of various behavioral indicators and the average displacement of the indicators when "boring" is pressed) -Σ (Correlation between the displacement of various physiological indices and the average displacement of the indices when "boring" is pressed) By setting the preference index, the subjective preference can be corrected by the behavioral index or physiological index. For example, even if a user presses "boring" in a quiz, if a physiological index (for example, electroencephalogram α2) shows a displacement close to that when "like" is pressed, the preference is corrected in the positive direction. Depending on the physiological index, there may be an index where there is almost no difference between the reaction when the "like" button is pressed and the reaction when the "boring" button is pressed. In that case, the same tendency is shown, so there is almost no deviation and the impact on the overall index is small. There are individual differences in which index has a relatively large impact. For some users, it may be prominent in the electroencephalogram, for others, it may be related to the heart rate, and for still other users, it may be related to the movement of the eye. Therefore, the correlation between each individual's preference (interest, concern) and the behavioral index or physiological index is analyzed, and the calculation method of the preference index is customized to suit the individual's characteristics. However, there are bound to be common instinctive reactions among humans, and if there is an index that shows the same tendencies across multiple user groups, the influence of that index will be increased, regardless of individual differences.

[0072] An example of a quiz difficulty-preference correlation diagram showing the correlation between quiz difficulty and preference will be described with reference to Fig. 7. The vertical axis of the diagram represents the preference for a quiz, that is, the degree of like or dislike or interest (positive is like, negative is dislike), and the horizontal axis represents the difficulty of the quiz (positive is difficult, negative is easy), and each quiz is plotted for each user.

[0073] When plotting the results of general users, that is, all users not bound by specific attributes, the first quadrant will be questions that bring back old memories, and the second quadrant will be questions that stimulate learning. For example, a quiz for fifth graders ("5th grade quiz" in the figure) is somewhat easy, but it is plotted in a somewhat interesting area because it brings back vague memories of what was learned long ago.

[0074] Quizzes about world heritage sites are interesting but have many somewhat difficult questions. Quizzes about famous paintings of the world are plotted as difficult but relatively interesting questions. Quizzes for first graders may be too easy and uninteresting. English quizzes for university entrance exams are more difficult, so they may become a genre that is uninteresting. Bar exam quizzes are full of extremely difficult questions, and users may not only lose interest, but dislike them. In this way, questions that are at least moderate in preference and of moderate difficulty, that is, not too difficult or too easy, are suitable for users to continue working on them with interest. The mapping of this quiz difficulty preference correlation diagram varies depending on the user.

[0075] An example of a quiz difficulty-preference correlation diagram by user attribute showing the correlation between quiz difficulty and preferences by user attribute will be described with reference to FIG.

[0076] A correlation diagram of quiz difficulty and preference for general people with no user attributes specified is expanded into a zero-dimensional map. Difficulty on the horizontal axis is plotted based on the average correct answer rate. Preference on the vertical axis is plotted based on the preference index. This makes it possible to create a collection of questions that everyone likes (is interested in) and that are neither too difficult nor too easy.

[0077] Similarly, a one-dimensional map in which one user attribute such as gender or age is selected can be expanded as many times as the number of attribute options. For example, by specifying an attribute such as "male" or "60s", a set of questions that are not too difficult or easy and that users with that attribute like (are interested in) can be created. Furthermore, a two-dimensional map can be created that combines multiple user attributes, such as "male" and "60s". In this way, by increasing the number of attributes to be combined, the number of samples that belong to the set that meets the conditions will be limited. By searching for a user group with more attributes, it is possible to identify a user group that is closer to the user's hobbies and tastes, and obtain a set of quiz questions of moderate difficulty that the user group likes. Ultimately, it is possible to drill down to the user's own quiz difficulty preference correlation diagram, but when the number of samples falls below a certain threshold, the number of applicable quiz questions will also be limited, so the compound conditions are relaxed according to the amount of data collected. If there is a sufficient amount of data from the user's personal data alone, questions to be asked may be selected from the user's quiz difficulty preference correlation diagram. If the amount of data is small for individual users, questions are selected from the quiz difficulty preference correlation diagram for a group of users with the same attributes. If the number of samples is insufficient, the number of dimensions is reduced to ensure the number of samples.

[0078] Figure 9 is an example of an attribute Venn diagram for an individual user. The numbers in circles in the diagram are represented here by numbers in parentheses. For example, if the user is a "male in his 60s who likes soccer," (1) if there is data for 100 general public people, then (2) 60 of them are male. (3) There are 20 people in their 60s, and (4) 12 of those in their 60s are male. Of those, (6) there are six people whose hobby is soccer. If the minimum sample number threshold is five, in this Venn diagram example, the sample set will be six people who are male in their 60s and have soccer as a hobby, with three attributes combined. If the minimum sample number threshold is raised to 10, the sample set will be 12 people who are male in their 60s and have two attributes combined, with two attributes combined, with two attributes combined, and one person combined. In this way, depending on the amount of data and the threshold settings, it is possible to obtain an appropriate number of samples even under complex conditions.

[0079] One way to find a group of users with similar hobbies and interests is to search for a group of users who share multiple matching attributes (for example, male, in his 30s, blood type O, from Tokyo, loves to travel, hobby is photography, etc.), but a group of users who don't have many things in common but share unique attributes is likely to have similar hobbies and interests in terms of sensibility. For example, 80% of the golfing population is male, but that doesn't mean that 80% of men play golf. The proportion of users who also have another attribute when one attribute is used as a parameter is called confidence.

[0080] Figure 10 shows an example of a user attribute Venn diagram (golf population). For example, if set X is male, set Y is female, and set Z is golfers, then confidence (X, Z) = σ(X∪Z) / σ(X)... the proportion of golfers among men, and confidence (Z, X) = σ(X∪Z) / σ(Z)... the proportion of men among golfers.

[0081] If we calculate these using Japan's population and golf population, we get confidence (X, Z) = 4.48 million / 61.02 million = 7.3%, and confidence (Z, X) = 4.48 million / 5.6 million = 80%.

[0082] There are 1.12 million women who play golf, which is only 1.7% of all women (64.48 million). In other words, the confidence level (Y, Z) is 1.7%, so although the number is small (minority), it can be said to be a very unique combination of attributes. The image of women who play golf is that they are sporty, celebrity-like, and fashion-conscious. In this way, even if you do not narrow it down by many combinations of attributes, if you can recognize the combination as unique, you can consider it as a group of users with the same hobbies and interests. However, even if the number of samples is small and it seems like a minority because there is not enough data collected, it cannot be said to be unique if the confidence level is high. The number of samples of users with the same hobbies and interests must be at least a certain number (for example, 10 cases), and similarly, it is defined as unique when the confidence level is below a certain number (for example, 5%).

[0083] This is a method of finding groups of users with the same hobbies and tastes by combining multiple attributes. Starting with the 0th group which targets the entire population and does not combine anything, the 1st group is narrowed down by one attribute, the 2nd group is narrowed down by a combination of two attributes, and so on up to the nth group, where the number of attribute combinations is n, the number of combinations is mCn=m! / n!(mn)!, where m is the maximum number of attributes, which is a fairly large number of combinations.

[0084] If m is 20 and n is 10, then the number of combinations becomes 184,756, which is a large number just to consider all the combinations of attributes, and if each attribute has an average of 5 possible options, the number of combinations becomes 5n (approximately 9.77 million if n=10), for a total of approximately 1.8 trillion combinations. Furthermore, while there is only one option for a single-choice attribute (e.g., gender), for a multiple-choice attribute (e.g., hobbies), if the number of options is x, then the number of variations increases by an additional 2x, making it unrealistic to analyze all the combinations.

[0085] Therefore, the following algorithm is used to reduce the number of cases. First, among all attributes, the attribute that has been selected by users the least (unique) is searched for. Those attributes that have been selected by fewer users than the minimum threshold minsup (e.g. 10) are excluded from the search. Conversely, if the number of users that have selected the attribute is greater than the maximum threshold maxsup (e.g. 100) and there is a sufficient population, the attribute is also excluded from the search for 0th order (single attribute with no combinations).

[0086] Although the total number of samples was small and fell below the maximum threshold, in order to exclude cases that cannot be considered unique, we calculate the expected value E of the number of occurrences when the occurrence of an item is completely random, and define it as unique if the number of occurrences is sufficiently small compared to the expected value (for example, within 10%). Let us show an easy-to-understand example. The probability of being male or female is about 50%, and the probability of playing golf or not is also 50% since there are two options. The expected value of a woman playing golf is 25%, but if the actual value is 1.7%, it is 1.7 / 25 = 6.8%, so it can be said to be sufficiently unique. However, in the case of a woman from Okinawa Prefecture, the expected value of being a woman is 50%, but the expected value of being from Okinawa Prefecture is 2.1%, since there are 47 prefectures to choose from. The expected value of a woman from Okinawa is about 1%. Even if the actual sample was 1.0.8%, it would be 0.8 / 1 = 80%, and it cannot be called unique.

[0087] Search for the least (unique) attribute among users who selected the same attribute as the user in the first stage (combination of two attributes). Those who selected the attribute with a minimum threshold (e.g. 10 cases) or less will be excluded. In addition, in the case of a combination of two or more attributes, the certainty is also checked. In the case of a combination of attribute X (e.g. female) and attribute Y (e.g. play golf), check both the certainty (X, Y) (e.g., certainty (female, play golf) is the percentage of women who play golf, 1.12 million / 64.48 million = about 1.7%) and the certainty (Y, X) (e.g., certainty (play golf, female) is the percentage of women who play golf, 1.12 million / 5.6 million = about 20%), and if either one is below a certain threshold (e.g. 5%), it is considered to be unique and is included in the target, but if both certainties are above a certain threshold, it is not considered to be unique and is excluded.

[0088] After checking all the first-order combinations, if a combination of attributes that is thought to be unique is found, a quiz difficulty preference correlation diagram is created for the group of users who have those attributes, and the quiz questions (quiz content) to be asked to the users are identified.

[0089] If the number of occurrences of the first combination is too large to identify a unique combination of attributes, a search is similarly performed for unique attributes in the second combination (three or more combinations).

[0090] Similarly, even if we proceed up to the nth order and try to search for unique attributes, the combination will end when the number of users for the (n+1)th newly combined attribute is less than the minimum threshold.

[0091] If the target user group is not found even after proceeding up to the nth order, return to the 0th order, increase the minimum threshold (for example, change it from 5% to 10%), and perform a similar search. The search continues until the target user group is identified or the minimum threshold is set to a relatively large value (for example, 50% of half). If the target user group cannot be identified even after the minimum threshold is set to a relatively large value, the Aproiri (frequent item set extraction) method, for example, is used. Details are omitted. If the target user group cannot be identified, non-hierarchical cluster analysis may be performed.

[0092] Figure 11 is an example of non-hierarchical cluster analysis (x-means) by user attributes. For example, the user's personal attributes are divided into clusters using the k-means method with a small number of clusters (for example, the total number of samples is divided into 5). The clusters generated in this way are divided using the 2-means method, and the Bayesian Information Criterion (BIC) is calculated. Division using the 2-means method is repeated until the BIC becomes larger than the previous BIC, searching for an appropriate value of x. When the BIC becomes large, the division is stopped at the previous division, and a user group that is the same as the cluster to which the user belongs is extracted as a result of the x-means non-hierarchical cluster analysis. Once a cluster (group of users) with the same interests and tastes is found, the user group is expanded into a quiz difficulty preference correlation diagram (or data) by user attribute. The method is to extract the user ID for each cluster (group of users) and obtain the values ​​in the user response database. The preference index and correct answer rate by quiz ID are tallied for all quiz IDs and all clusters (users).

[0093] A group of quizzes with a preference index above a certain threshold (like) and of moderate difficulty, that is, quizzes in a genre that is liked by all users and that are neither too difficult nor too easy, are selected and presented.

[0094] If the quiz is presented multiple times and the user response continues to be poor at a certain rate or more, selection from that quiz group is discontinued and selection is restarted from another user group with similar hobbies and tastes.

[0095] Using FIG. 12, examples of quiz screens G11 to G13 provided to the user from the user terminal 1 will be described. As shown in the screen G11, for example, a quiz about world paintings may be considered, in which photos of famous paintings are displayed and the quiz asks the user to guess the artist. The answer is in a multiple choice format, and for example, the user can answer one of the four options from 1 to 4 by saying "number one" out loud, or by touching the option "(1)" on the touch screen. An option not to answer is also provided. In that case, the user answers by touching the "Skip: too easy", "Skip: too difficult", or "Skip: not interested" button together with the reason why the user did not answer, or by verbally saying "too easy", "too difficult", "not interested", or simply "Skip".

[0096] When the user selects one of the options (1) to (4), the answer (correct or incorrect) and a brief explanation are displayed on the screen G12 and are also output as audio. The user can increase his / her knowledge by listening to the brief explanation. If the user is further interested, he / she can listen to a detailed explanation that is prepared in advance for some quiz questions. If the user finds the quiz interesting but not interesting enough to listen to a detailed explanation, he / she presses the "Like" button or says "Like". If the user finds the brief explanation boring after listening to it, he / she presses the "Boring" button or says "Boring". If he / she cannot answer either way, he / she presses the "Next" button or says "Next".

[0097] When the More Detailed Explanation button is pressed, a more detailed explanation screen is displayed on the user terminal 1 as shown in screen G13, and an audio explanation is output. If the user is further interested, the user can press the "Anticipated Question / Trivia" button or say "Anticipated Question" or "Trivia" to be provided with some mini-knowledge related to the quiz.

[0098] If multiple pieces of trivia are available, an index screen (not shown) introducing the trivia is displayed. The user can view the trivia they want to know by pressing the number of the "anticipated question / trivia" they want to know about or by speaking the number. After listening to one piece of prepared trivia, they can move on to the next piece of trivia or go back to the previous piece of trivia. By pressing the back button, they can return to the index display screen and jump to the playback of their favorite trivia. Since a more detailed explanation or a more frequent playback of a trivia indicates that the user was interested in the quiz question, the explanation click rate and trivia click rate, which are behavioral indicators in the user response database, are updated.

[0099] The overall processing of the knowledge information providing system 3 will be described with reference to the flowchart of Fig. 13. Various assessments are carried out to obtain initial values ​​related to the attributes of the user (S10). The details of the assessments will be described later.

[0100] By repeating the assessment, question setting, and answering, the user reaction database 262 is created. The question setting control unit 21 selects a quiz with a high level of preference (level of interest and concern) and a medium level of difficulty based on the user's subjective index from the data accumulated in the user reaction database 262 (S11). As long as there is a corresponding quiz, the question setting control unit 21 sets that question (S12: YES, S16). When there are no more quizzes that can be extracted based on the subjective index (S13: NO), the question setting control unit 21 proceeds to step S13.

[0101] The question setting control unit 21 selects a quiz with a high preference (interest level) and a medium level of difficulty based on the user's behavioral and physiological indices from the user response database 262 (S13). If there is a question that was not extracted from the subjective index alone but which clearly shows that the user's interest has increased based on the behavioral and physiological indices, a question from that genre is set. As long as there is a suitable quiz, questions from that group are set (S14: YES, S16). When there are no more suitable quizzes (S14: NO), the question setting control unit 21 proceeds to step S15.

[0102] The question setting control unit 21 selects a quiz with a high preference (interest level, concern level) and a medium level of difficulty from the difficulty-preference correlation diagram of the user group belonging to the same cluster (S15). If there is no corresponding quiz, the difficulty threshold may be increased. If a corresponding quiz is still not found, the question setting control unit 21 can also increase the preference threshold to search for and set a corresponding quiz.

[0103] The question setting control unit 21 randomly sets a question from among the selected quizzes (S16). The question setting control unit 21 determines whether the user's answer received from the user terminal 1 is correct or incorrect (S17). The question setting control unit 21 registers the correct or incorrect answer of the user in the user response database 262. If the user selects "skip", the question setting control unit 21 adjusts the difficulty level depending on the reason for skipping (S18-S20).

[0104] That is, if the user feels that the quiz is "too difficult" (S18: difficult), the question setting control unit 21 lowers the threshold for the percentage of correct answers and causes the user to reselect the quiz (S20, S16). Conversely, if the user feels that the quiz is too easy (S18: easy), the question setting control unit 21 raises the threshold for the percentage of correct answers and selects the quiz (S19, S16). If the quiz does not interest the user, the question setting control unit 21 records in the user reaction database 262 that the quiz (the question set) did not interest the user (S21). The question setting control unit 21 repeats the above process until the user finishes using the knowledge information provision system 1 (S22).

[0105] The assessment process (S10) in Fig. 13 will be described in detail with reference to the flowchart in Fig. 14. The assessment includes an individual assessment and a quiz assessment.

[0106] The question setting control unit 21 checks whether the user has already taken the individual assessment (S100). If the individual assessment has not been taken (S100: NO), the question setting control unit 21 proceeds to individual assessment processing (S101) which will be described later with reference to FIG.

[0107] The question setting control unit 21 checks whether the quiz assessment has already been carried out (S102), and if not (S102: NO), the process proceeds to a quiz assessment process (S103) which will be described later with reference to FIG.

[0108] The question setting control unit 21 determines whether a predetermined amount of assessment information has been accumulated in steps S101 and S103 (S104), and if it determines that it has been accumulated (S104: YES), it executes an apoori analysis (S105) and a non-hierarchical cluster analysis (S106), and ends this process. If the predetermined amount of assessment information has not been obtained (S104: NO), the question setting control unit 21 ends this process. The assessment process is performed on another occasion, and assessment information about the user is accumulated.

[0109] The individual assessment process (S101) will be described in detail with reference to the flowchart in Fig. 15. When a user uses the knowledge information provision system 3 for the first time, the question setting control unit 21 collects information on individual attributes so as to set a more suitable quiz.

[0110] The question setting control unit 21 selects an item that has not yet been set in the personal attribute database 261 (S1010), and asks the user about the selected item from the user terminal 1 (S1011). When the question setting control unit 21 receives an answer to the question from the user terminal 1, it writes the answer in the personal attribute database 261 (S1012). The question setting control unit 21 repeatedly executes steps S1010-S1012 until all items for the user have been checked (S1013).

[0111] The quiz assessment process (S103) will be described in detail with reference to the flowchart in Fig. 16. As with the personal assessment, when a user first uses the knowledge information providing system 3, the question setting control unit 21 sets a sample quiz so as to set a more suitable quiz, and collects the user's quiz difficulty preference information.

[0112] The question setting control unit 21 randomly selects one question from the quizzes stored in the quiz database 25 (S1030), and sets the question to the user from the user terminal 1 (S1031). The question setting control unit 21 receives the user's answer and determines whether it is correct or not (S1032).

[0113] If the user's answer is correct or incorrect, the question setting control unit 21 updates the user response database 262 (S1036). If the user's answer is "skip" (S1033), the question setting control unit 21 adjusts the difficulty of the next quiz depending on the reason why the user skipped the quiz (S1034, S1035). If the user feels that the quiz is easy, the question setting control unit 21 increases the difficulty of the next quiz (S1034). If the user determines that the quiz is difficult, the question setting control unit 21 lowers the difficulty of the next quiz (S1035).

[0114] The question setting control unit 21 judges whether the number of sample data obtained in the quiz assessment reaches a predetermined threshold ThS (S1037), and if it judges that it has reached the threshold (S1037: YES), the process ends. If it judges that it has not reached the threshold (S1037: NO), the question setting control unit 21 randomly selects a quiz of another genre (S1038) and presents the selected quiz to the user (S1031).

[0115] The process (S21) of updating the user reaction database 262 will be described with reference to the flowchart of Fig. 17. After presenting a quiz, including a quiz assessment, to the user, the question setting control unit 21 receives the user's reaction and updates the user reaction database.

[0116] The question setting control unit 21 registers subjective index data, which is information on the button pressed by the user (such as the "Like" button or the "Boring" button), in the user response database 262 (S210). The question setting control unit 21 registers user behavior index data, such as the time it took the user to answer and the accuracy rate, in the user response database 262 (S211). The question setting control unit 21 registers physiological index data, such as brain waves, heart rate, and respiratory rate, that were observed at that time, in the user response database 262 (S212). The question setting control unit 21 calculates a preference (interest, curiosity) index from these indices, and registers it in the user response database 262 (S213).

[0117] The process of posing a quiz to a user (S300) will be described with reference to the flowchart in Fig. 18. The method of posing a quiz to a user is basically based on the user's own subjective opinion, which determines whether the user likes or dislikes something.

[0118] That is, the question setting control unit 21 calculates a preference index from the subjective index, and sets a certain number of questions from a question genre with a preference level or higher and a group of questions with a moderate correct answer rate for the user (S301). If the answer result is a lot of "Like" clicks, that is, if the "Like" rate is equal to or higher than a certain threshold Th1, it is determined that the user likes questions of that level in that field, and questions are continued to be set based on the same subjective index (S302: YES).

[0119] If the like rate of the user falls below the threshold Th1 (S302: NO), the question setting control unit 21 sets quizzes based on the behavioral index and physiological index (S303) since it is necessary to change the question setting tendency. If the like rate is equal to or greater than the threshold Th1 (S304: YES), the question setting control unit 21 repeats step S303 and sets quizzes based on the behavioral index and physiological index of the user.

[0120] If the "Like" rate does not reach or exceed the threshold Th1 after a certain number of quizzes have been set (S304: NO), a group of users with similar interests and tastes is identified, and quizzes are set based on the quiz difficulty preference correlation diagram of the user group, rather than just the characteristics of the individual user (S305). If the "Like" rate does not reach or exceed the threshold Th1 after a certain number of quizzes have been set (S306: NO), the user reaction database 262 is updated in the above-mentioned manner (S21), and unless the user stops using the knowledge information provision system 3 (S307: NO), the process returns to step S301 to set quizzes.

[0121] We will now explain the quiz categorization mechanism. Some quiz questions do not have a defined genre or subcategory, or are incorrectly categorized. Therefore, we perform natural language analysis on the text written in the quiz question, sample answers, explanations, trivia information, etc., and categorize the questions in two ways. The first method is (1) to apply the questions to existing classifications. The second method is (2) to create new classification criteria.

[0122] First, (1) fitting into existing classifications is a method of classifying various sentences related to the quiz questions into words using morphological analysis, investigating the number of occurrences (frequency) of nouns and proper nouns using the bag of words method, and determining the best existing genre (category) and subcategory to classify them into. The number of occurrences can be used as a feature to analyze using neural network techniques, or classification can be performed using classifiers such as Naive Bayes, SVM (Support Vector Machine), and logit models, which are traditional machine learning methods. Furthermore, a classification can be determined based on the TF-IDF (Term Frequency - Inverse Document Frequency) method, which is an index obtained by dividing the number of occurrences of a word by the proportion of the total number of documents in which the word appears.

[0123] The second point (2) is to create new classification criteria. In the case of a question in (1) in which only words from existing classification categories appear at a frequency below a certain threshold, it may be inappropriate to classify the question into an existing genre or subcategory. In this case, it is possible to analyze the text of the quiz question and generate a new genre or subcategory.

[0124] According to this embodiment configured in this manner, the next knowledge information can be provided to the user terminal in accordance with the user's reaction to the knowledge information and predetermined conditions, thereby enabling the user to intellectually activate his or her leisure time.

[0125] According to this embodiment, knowledge information suited to a user can be provided based on at least one of a subjective index relating to a predetermined attribute of the user, a behavioral index relating to the user's behavior in response to a quiz, and a physiological index relating to predetermined vital data of the user.

[0126] According to this embodiment, knowledge information is provided to a user based on a subjective index regarding the user's attributes and the responses of other users in the user attribute cluster (user group) to which the user belongs, so that even if there is little information (assessment information) about the user, knowledge information suitable for the user can be provided.

[0127] According to this embodiment, first, knowledge information is selected based on a subjective index, and second, knowledge information is selected based on a behavioral index and a physiological index, so that knowledge information can be provided first based on the emotion directly expressed by the user, and then based on the user's latent emotion. Therefore, knowledge information can be provided to the user taking into consideration not only the user's overt emotion but also his / her hidden emotion. The present invention is not limited to the above-described embodiment. A person skilled in the art can make various additions and modifications within the scope of the present invention. The above-described embodiment is not limited to the configuration example shown in the attached drawings. The configuration and processing method of the embodiment can be appropriately changed within the scope of achieving the object of the present invention.

[0128] In addition, the components of the present invention may be selected arbitrarily, and the invention having the selected components is also included in the present invention. Furthermore, the components described in the claims may be combined in combinations other than those explicitly stated in the claims. [Explanation of symbols]

[0129] 1: user terminal, 2: server, 3: knowledge information providing system, 21: question setting control unit, 22: data transmission / reception unit, 23: user attribute analysis unit, 24: user response analysis unit, 25: quiz database, 26: user information database

Claims

1. A knowledge information providing system comprising a server and at least one user terminal connected to the server via a communication network, the system providing digitized knowledge information from the server to the user terminal, The server transmits the knowledge information selected for each user according to a predetermined condition to the user terminal, acquires a user's reaction to the knowledge information from the user terminal, and provides the next knowledge information to the user terminal according to the acquired user's reaction and the predetermined condition. Knowledge information provision system.

2. The knowledge information is information including a quiz, and the user's reaction includes at least the user's answer to the quiz, the user's impression of the quiz, and the user's physiological reaction. The knowledge information providing system according to claim 1 .

3. The predetermined condition is set based on a subjective index regarding a predetermined attribute of the user. The knowledge information providing system according to claim 2.

4. The predetermined condition is further set based on a physiological index related to predetermined vital data of the user. The knowledge information providing system according to claim 3.

5. The predetermined condition is further set based on a behavior index related to the user's behavior with respect to the quiz. The knowledge information providing system according to claim 4.

6. The predetermined condition is set based on at least one or more of a subjective index related to a predetermined attribute of the user, a physiological index related to a predetermined vital data of the user, and a behavioral index related to a behavior of the user in response to the quiz. The knowledge information providing system according to claim 2.

7. The predetermined condition is set based on a subjective index regarding the user's attributes and the responses of other users in the user attribute cluster to which the user belongs. The knowledge information providing system according to claim 2.

8. The predetermined conditions are, first, selecting knowledge information based on the subjective index, and, second, selecting knowledge information based on the behavioral index and the physiological index. The knowledge information providing system according to claim 6.

9. A method for providing knowledge information, comprising: a server; and at least one user terminal connected to the server via a communication network, the method providing digitized knowledge information from the server to the user terminal, the method comprising: The server, Transmitting the knowledge information selected for each user according to a predetermined condition to the user terminal; acquiring a user's reaction to the knowledge information from the user terminal; Providing next knowledge information to the user terminal according to the acquired user reaction and the predetermined condition. Knowledge information provision method.

Images