Daily report analysis apparatus, daily report analysis method, and daily report analysis program

Abstract

To propose a daily report analysis apparatus, a daily report analysis method, and a daily report analysis program for picking out cases that require attention and prevent case failures before they occur.SOLUTION: A daily report analysis apparatus for outputting priorities of daily reports to be confirmed from daily reports of a plurality of cases, comprises: a storage device that stores a daily report analysis program for determining the priorities of daily reports of the plurality of cases and outputting the information of the priorities; and a processor that reads and executes the daily report analysis program from the storage device. The processor converts the text information of the daily report into a text vector represented by a numeric vector, calculates an evaluation value for each preset evaluation index, calculates a progress deviation based on the evaluation value of each evaluation index and a progress registration value reported on the daily report, calculates a confirmation priority indicating the confirmation priority for each case based on the progress deviation of the plurality of cases obtained by the deviation calculation process; and visualizes the plurality of cases by arranging them based on a numeric number of the confirmation priority.SELECTED DRAWING: Figure 8

Description

[Technical Field]

[0001] The present invention relates to a daily report analysis device, a daily report analysis method, and a daily report analysis program. [Background technology]

[0002] In corporate project activities (sales activities or product manufacturing activities at a factory), the person in charge may write down the details of the activities in text (hereafter referred to as text information) in daily reports or other reporting documents, and report the daily reports and project progress to the manager. The manager judges the progress from the text information based on the evaluation indicators used to assess progress, and checks to see if there is any discrepancy between the progress judged by the manager and the progress reported by the person in charge. If there is a discrepancy in progress, a misunderstanding may arise between the parties involved in the project, which could affect future activities. Therefore, it is important from the perspective of project management for the manager to check the text information in the daily reports.

[0003] However, since managers have many cases to manage and reading each piece of text information takes time, the workload on managers is heavy. For this reason, in recent years, there have been an increasing number of cases where text information is analyzed using AI technology to improve the efficiency of case management.

[0004] For example, Patent Document 1 discloses a technique for analyzing the text of daily reports, in which the presence or absence of pre-set risk-related keywords is determined from the text information of the daily sales reports, and the risk is grasped. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2020-119174 [Non-patent literature]

[0006] [Non-Patent Document 1] Jacob Devlin, et al., "Bert: Pre-training of deep bidirectional transformers for language under-standing." Proceedings of naacL-HLT. Vol. 1. (2019) 1810.04805v2.pdf (arxiv.org) [Non-patent document 2] Nils Reimers and Iryna Gurevych, “Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks”, arXiv preprint arXiv:1908.10084 (2019) https: / / arxiv.org / pdf / 1908.10084.pdf Summary of the Invention [Problem to be solved by the invention]

[0007] However, with conventional technologies such as that disclosed in Patent Document 1, even if there is a discrepancy between the progress (estimated progress) that can be estimated from the information written in a daily report (sales information in the case of sales activities) and the progress (registered progress) that a person in charge has registered in the daily report, the discrepancy cannot be perceived as a risk. For example, taking sales activities as an example, discrepancies between estimated progress and registered progress can occur for various reasons, including: (i) the lack of established know-how appropriate to the situation of sales activities means that the progress that should be registered is left to the personal judgment of the person in charge (personalization of sales activities), (ii) in order to avoid attracting attention, sales people may intentionally enter progress (registered progress) in their daily reports that differs from the progress that can be extracted from the sales information (estimated progress) (deception), and (iii) the daily reports may not adequately include information that serves as the basis for determining estimated progress, such as budgets and needs (insufficient information).

[0008] As mentioned above, managers often have many projects to manage. Therefore, they must prioritize checking the daily reports of important projects and move them forward. However, due to the manager's subjective judgment, situations can arise where they do not check the daily reports of projects that should be prioritized, resulting in the project failing. The present invention has been made in view of such circumstances, and proposes a daily report analysis technique that picks out cases that require attention and prevents case failures from occurring. [Means for solving the problem]

[0009] In order to solve the above problems, the present invention proposes a daily report analysis device that outputs the priority of daily reports to be confirmed from the daily reports of multiple cases, and includes: a storage device that stores a daily report analysis program for determining the priority of the daily reports of multiple cases and outputting information about the priority; and a processor that reads and executes the daily report analysis program from the storage device. The processor performs the following steps: preprocessing to convert the text information of the daily reports into text vectors represented by numerical vectors; evaluation value calculation processing that uses the text vectors to calculate evaluation values ​​for each predetermined evaluation index; deviation calculation processing that calculates a progress deviation using the evaluation values ​​of each evaluation index and the progress registration value reported in the daily report; confirmation priority calculation processing that calculates a confirmation priority indicating the confirmation priority for each case based on the progress deviation of the multiple cases obtained by the deviation calculation processing; and output processing that arranges and visualizes the multiple cases based on the numerical values ​​of the confirmation priority.

[0010] Further features related to the present invention will become apparent from the description and accompanying drawings of this specification, and aspects of the present invention may be realized and realized by the elements and combinations of various elements and aspects set forth in the following detailed description and the appended claims. It should be understood that the descriptions in this specification are exemplary and illustrative only and are not intended to limit the scope or application of the claims in any way. [Effects of the Invention]

[0011] The daily report analysis technique of the present invention makes it possible to pick out cases that require attention and prevent failure of the cases before they occur. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram showing an example of a schematic configuration of a daily report analysis device 1 according to an embodiment of the present invention. [Figure 2] FIG. 10 is a diagram showing an example of the configuration of a daily report management table 151. [Figure 3] FIG. 10 is a diagram illustrating an example of the configuration of an evaluation index setting table. [Figure 4] FIG. 10 is a diagram illustrating an example of the configuration of a progress setting table 153. [Figure 5A] FIG. 2 is a diagram illustrating a first example of a software configuration for daily report analysis processing according to the present embodiment. [Figure 5B] FIG. 10 is a diagram illustrating a second example of a software configuration for daily report analysis processing according to the present embodiment. [Figure 5C] 10 is a table showing examples of set target values ​​for evaluation values. [Figure 6] FIG. 10 is a diagram illustrating an example of a list visualized in descending order of confirmation priority. [Figure 7] 10 is a flowchart for explaining an outline of a daily report analysis process according to the present embodiment. [Figure 8] 10 is a flowchart illustrating details of daily report analysis processing (text analysis processing). [Figure 9] FIG. 10 is a diagram illustrating an example of the software configuration of a daily report analysis process including an evaluation value calculation process 52 according to a first modified example. [Figure 10] 10 is a diagram in which information generated by each process in the description determination process 91, the updated evaluation value calculation process 92, and the weighted average process 93 that make up the evaluation value calculation process 52 shown in FIG. 9 is added. [Figure 11] FIG. 10 is a diagram showing an example software configuration (variation 2) of daily report analysis processing having feature extraction processing 111 as an auxiliary function. [Figure 12]FIG. 10 is a diagram showing a display example (Modification 3) in which the case confirmation priority is visualized using a two-axis graph of progress deviation and case importance. DETAILED DESCRIPTION OF THE INVENTION

[0013] This embodiment relates to a technology for calculating confirmation priorities, which are the confirmation priorities for multiple cases, based on the text information in the daily report for each case (e.g., sales case) and the progress information (progress registration values) registered by the person in charge (sales representative).

[0014] Hereinafter, the present embodiment will be described with reference to the accompanying drawings. In the accompanying drawings, functionally identical elements may be denoted by the same numerals, and redundant explanations will be omitted. Note that the accompanying drawings show specific embodiments and examples in accordance with the principles of the present disclosure, but these are intended to aid in understanding the present disclosure and are by no means to be used to interpret the present disclosure in a limiting manner.

[0015] Although the present embodiment has been described in sufficient detail to enable those skilled in the art to implement the present disclosure, it should be understood that other implementations and forms are possible, and that changes in configuration and structure and substitutions of various elements are possible without departing from the scope and spirit of the technical ideas of the present disclosure. Therefore, the following description should not be interpreted as being limited thereto.

[0016] <Example of daily report analyzer configuration> 1 is a diagram showing an example of a schematic configuration of a daily report analysis device 1 according to an embodiment of the present invention. The daily report analysis device 1 includes, for example, an input device 11, an output device 12, a central processing unit (processor) 13, a main memory device 14, and an auxiliary memory device 15.

[0017] The main memory device 14 stores a program (daily report analysis processing program) for executing the daily report analysis processing according to this embodiment. The central processing unit 13 reads the daily report analysis processing program stored in the main memory device 14 and deploys it in internal memory to construct a daily report analysis processing unit. The auxiliary memory device 15 stores data necessary for executing the daily report analysis processing. The input device 11 and output device 12 input and output data.

[0018] The auxiliary storage device 15 stores, for example, a daily report management table 151, an evaluation index setting table 152, and a progress setting table 153. Before the daily report analysis process is executed by the daily report analysis device 1, the input device 11 is used to store necessary data in each table.

[0019] <Contents of various tables> (i) Daily Report Management Table 151 FIG. 2 is a diagram showing an example of the configuration of the daily report management table 151. The daily report management table 151 stores data on reports such as daily reports used by a company, and has as its configuration information a case ID 1511 for uniquely identifying a case, a report date 1512, report content 1513, and case progress 1514. Using the case ID 1511 as a key, information such as the report date 1512, report content (text information) 1513, and case progress 1514 is set in the daily report management table 151. Here, the case ID 1511 is a required field. The report date 1512 manages the time series in which daily reports for each case are registered. The third and subsequent columns are used as input values. In addition to these pieces of data, the daily report management table 151 may also store various other items.

[0020] (ii) Evaluation index setting table 152 FIG. 3 is a diagram showing an example of the configuration of an evaluation index setting table. The evaluation index setting table 152 stores the items and evaluation values ​​of evaluation indexes for evaluating progress. The first row of the evaluation index setting table 152 registers each evaluation index item such as budget, needs, and implementation time. The second and subsequent rows of the evaluation index setting table 152 register the score for each index and the content corresponding to the score. Here, the domain of the evaluation value does not need to be uniform for each index. For example, the registered content may be a four-level rating for budget and a two-level rating for needs.

[0021] (iii) Progress setting table 153 4 is a diagram showing an example of the configuration of the progress setting table 153. The progress setting table 153 stores a plurality of steps (progress) that indicate the stage (status) at which the case is. In the progress setting table 153, the first line describes the numerical value of the progress status (the larger the numerical value, the more progress there is), and the second line describes the status corresponding to the numerical value.

[0022] <Example of software configuration for daily report analysis processing> 5A is a diagram showing an example of the software configuration of the daily report analysis process according to this embodiment. The daily report analysis process is made up of a pre-processing step 51, an evaluation value calculation step 52, a deviation calculation step 53, and a confirmation priority calculation step 54.

[0023] Pre-processing 51 is a process for converting text information of daily reports (information input by the user or information obtained from the daily report management table 151) into numerical vectors (hereinafter referred to as text vectors). Conversion methods include, for example, techniques such as BERT (see Non-Patent Document 1). However, the user can freely select which method to use.

[0024] The evaluation value calculation process 52 is a process for calculating an evaluation value for each evaluation index using the text vectors generated by the preprocessing 51. For example, the evaluation value is calculated as follows: budget is 1 point (not implemented), needs is 3 points (there are serious issues), and implementation time is 2 points (within 2 years). The calculation method includes, for example, techniques such as machine learning and deep learning. However, the user can freely select which method to use.

[0025] The deviation calculation process 53 is a process that uses machine learning or deep learning to calculate a progress estimate value from the evaluation value of each evaluation item (each evaluation index), and calculates a progress deviation degree based on the progress estimate value and the progress registration value (progress reported in the daily report). The progress deviation degree can be calculated, for example, by differential calculation. However, the user can arbitrarily select which method to use (see FIG. 5A).

[0026] Alternatively, as shown in FIG. 5B, the deviation calculation process 53 may use the target evaluation value set by the user (e.g., a manager) for each progress step using the target evaluation value setting unit 55 (e.g., set target evaluation values ​​are budget 1, needs 2, and implementation time 2; see "Hearing" in FIG. 5C) and the evaluation value of each of the above evaluation items (e.g., the calculated evaluation values ​​for each evaluation item are budget 1, needs 1, and implementation time 1), to calculate the difference between the evaluation value of each of the above evaluation items that is below the target evaluation value (e.g., deviation for each item: budget 0, needs 1, and implementation time 1), and generate a deviation (in this example, the deviation is 2).

[0027] The confirmation priority calculation process 54 is a process that acquires the progress deviations of multiple cases obtained by the deviation calculation process 53, calculates the confirmation priority, which is the priority order for confirmation of the cases, and displays it to the user. One display method is to visualize the confirmation priority in a list in descending order of priority. However, the user can freely select which method to use.

[0028] Figure 6 shows an example of the structure of a visualized list sorted in descending order of confirmation priority. By outputting such a list, the user (manager) can easily find out which projects have a large discrepancy between estimated progress and registered progress, without having to check the daily reports in detail, and can identify which projects should be rechecked or given priority.

[0029] <Outline of daily report analysis processing> FIG. 7 is a flowchart for explaining an outline of the daily report analysis process according to this embodiment.

[0030] (i) Step S701 The daily report analysis processing unit generated in the central processing unit 13 acquires daily report data input using the input device 11. The daily report data stored in the daily report management table 151 may be input to the daily report analysis processing unit in response to a user instruction.

[0031] (ii) Step S702 The daily report analysis processing unit calculates confirmation priorities from the text information of the daily reports of multiple cases and the progress registration values, and displays the results.

[0032] <Details of daily report analysis processing> The daily report analysis process can also be called text analysis process, and as explained in the software configuration above, it is a process that calculates the confirmation priority from the text information of the input daily report and the progress registration value (progress recorded in the daily report). Figure 8 is a flowchart for explaining the daily report analysis process (text analysis process) in detail.

[0033] (i) Step S801 The daily report analysis processing unit converts the text information of the daily report (information input by the user or information obtained from the daily report management table 151) into a numerical vector (hereinafter referred to as a text vector) (corresponding to pre-processing 51). For example, there is a technique such as BERT (see Non-Patent Document 1) as a conversion method. However, the user can freely select which method to use.

[0034] (ii) Step S802 The daily report analysis processing unit calculates an evaluation value for each evaluation index using the text vectors generated in step S801 (corresponding to evaluation value calculation process 52). For example, the evaluation value is calculated as follows: budget is 1 point (not implemented), needs is 3 points (there are serious issues), and implementation time is 2 points (within 2 years). Calculation methods include, for example, machine learning and deep learning techniques. However, the user can freely select which method to use.

[0035] (iii) Step S803 The daily report analysis processing unit uses machine learning or deep learning to calculate a progress estimate from the evaluation value of each of the evaluation items (each evaluation index), and calculates a progress deviation using the progress estimate and the progress registration value (progress reported in the daily report) (corresponding to the process of deviation calculation process 53 in FIG. 5A). Methods for calculating the progress deviation include, for example, differential calculation. However, the user can freely select which method to use.

[0036] Alternatively, the daily report analysis processing unit user (e.g., a manager) may use the target evaluation value set by the daily report analysis processing unit user (e.g., a manager) for each progress in the target evaluation value setting unit 55 (e.g., set target evaluation values ​​are budget 1, needs 2, and implementation time 2; see "Hearing" in Figure 5C) and the evaluation value of each of the above evaluation items (e.g., the calculated evaluation values ​​of each evaluation item are budget 1, needs 1, and implementation time 1) to calculate the difference between the evaluation value of each of the above evaluation items that is below the target evaluation value (e.g., deviation for each item: budget 0, needs 1, and implementation time 1), and generate a deviation (in this example, deviation 2) (equivalent to the processing of deviation calculation process 53 in Figure 5B).

[0037] (iv) Step S804 The daily report analysis processing unit acquires the progress deviations of the multiple cases obtained in step S803, calculates the confirmation priority, which is the priority order for confirmation of the cases, and displays it to the user (corresponding to confirmation priority calculation process 54). One display method is to visualize the results in a list in descending order of confirmation priority. However, the user can freely select which method to use.

[0038] By sorting the cases in descending order of confirmation priority, the user (manager) can easily find out which cases have a large discrepancy between estimated progress and registered progress without having to check the daily reports in detail, and can identify which cases should be rechecked or given priority.

[0039] <Variation 1> 9 is a diagram showing an example of the software configuration of the daily report analysis process including the evaluation value calculation process 52 according to Modification 1. The pre-processing 51, the deviation calculation process 53, and the confirmation priority calculation process 54 are as described above, and therefore detailed explanations thereof will be omitted here.

[0040] The evaluation value calculation process 52 according to the first modification calculates an evaluation value (post-update evaluation value) for each indicator in the daily report to be analyzed based on the text vectors generated by the pre-processing 51, and calculates an evaluation value for each evaluation indicator (evaluation item) in the daily report to be analyzed using this post-update evaluation value and the evaluation value before the daily report to be analyzed was registered (pre-update evaluation value). For one project, a daily report is created each time an activity is carried out, and daily report data is accumulated in chronological order in the daily report management table 151, but information for each evaluation indicator is not necessarily updated every time. Therefore, for indicators that have not been updated, the final evaluation value is calculated taking into account the evaluation value in the daily report before the update.

[0041] In the evaluation value calculation process 52, the daily report analysis processing unit calculates an updated evaluation value for the currently input daily report in the updated evaluation value calculation process 92 based on the text vector obtained in the pre-processing 51.

[0042] Furthermore, in the description determination process 91, the daily report analysis processing unit calculates the description probability p (p is a real number between 0 and 1) using the text vector obtained in the preprocessing 51. Note that since actual daily reports often contain ambiguous expressions, the description determination process 91 is a process that expresses, in terms of probability, how clearly the evaluation indexes (evaluation items) are written in the daily report.

[0043] Furthermore, in the weighted average process 93, the daily report analysis processor calculates a weighted average of the post-update evaluation value and the pre-update evaluation value, taking into account the description probability p, and outputs the result as an evaluation value. In other words, evaluation value = p × post-update evaluation value + (1-p) × pre-update evaluation value.

[0044] FIG. 10 is a diagram in which information generated by each of the description determination process 91, updated evaluation value calculation process 92, and weighted average process 93, which constitute the evaluation value calculation process 52 shown in FIG. 9, is added. An example is shown in which it is determined that the budget is not described and the needs are described. In this case, the budget evaluation value is output as the pre-update evaluation value, and the needs evaluation value is output by the evaluation value calculation unit. Note that at least one of the updated evaluation value calculation process 92 or the description determination process 91 can use machine learning or deep learning. For example, the description probability p can be calculated using a sigmoid function in machine learning or deep learning. However, the user can freely select the method to be used in the updated evaluation value calculation process 92 and the description determination process 91.

[0045] <Variation 2> 11 is a diagram showing an example of the software configuration of daily report analysis processing having a feature extraction process 111 as an auxiliary function. The feature extraction process 111 is an auxiliary function used during learning of the evaluation value calculation process 52 when the evaluation value calculation process 52 is executed by applying machine learning or deep learning.

[0046] The feature extraction process 111 is a process for calculating (extracting) a feature vector of an evaluation value using the text vector obtained by the preprocessing 51. When training a machine learning or deep learning model, if the training data (evaluation value training data) for the evaluation values ​​of two different text vectors is the same, the model is trained to increase the similarity between the two feature vectors. If the training data is different, the model is trained to decrease the similarity between the two feature vectors. By using the feature extraction process 111, the evaluation value calculation process 52 can be predicted with high accuracy even when there is little training data. Details of the feature extraction process 111 are described, for example, in Non-Patent Document 2. Examples of methods for calculating similarity include cosine similarity and Euclidean distance. However, the user can freely select which method to use to calculate similarity. The feature extraction process 111 is also a technique that can be used when configuring the description determination process 91 using machine learning or deep learning.

[0047] <Variation 3> In the daily report analysis process, the confirmation priority calculation process 54 may calculate the confirmation priority taking into consideration not only the progress deviation but also the importance of the item.

[0048] If the user (manager) sets the case importance in advance, which is a value indicating the importance of each case, the daily report analysis processor can calculate the case confirmation level using the progress deviation and case importance. Examples of case importance include whether the customer is important or the size of the order amount. Here, multiple case importance items may be set.

[0049] When using project importance, the project confirmation priority can be calculated by, for example, taking the sum of the project importance and the deviation. Furthermore, one way to display this to the user is to visualize it as a two-axis graph, as shown in Figure 12. That is, it is possible to output multiple types of project confirmation priority displays, such as progress deviation and customer importance, and progress deviation and order amount. In the case of Figure 12, it can be seen that the project confirmation priority is higher the further to the top right of the graph. However, the user can freely select which display method to use.

[0050] <Other> The functions of the present embodiment can also be realized by software program code. In this case, a storage medium on which the program code is recorded is provided to a system or device, and the computer (or CPU or MPU) of the system or device reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the program code itself and the storage medium on which it is stored constitute the present disclosure. Examples of storage media for providing such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, and ROMs.

[0051] In addition, an operating system (OS) running on a computer may perform some or all of the actual processing based on instructions in the program code, and the functions of the above-described embodiments may be realized by this processing. Furthermore, after the program code is read from a storage medium and written to a memory on a computer, a CPU of the computer may perform some or all of the actual processing based on instructions in the program code, and the functions of the above-described embodiments may be realized by this processing.

[0052] Furthermore, the program code of the software that realizes the functions of the embodiments may be distributed via a network and stored in a storage means such as a hard disk or memory of the system or device, or in a storage medium such as a CD-RW or CD-R, so that when in use, the computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage means or storage medium.

[0053] The processes and techniques described herein are not inherently related to any specific device and can be implemented by a combination of components. Various types of general-purpose devices can also be added. A dedicated device may be constructed to perform the functions of the present embodiment. Various functions can also be formed by appropriately combining multiple components disclosed in the present embodiment. For example, some components may be omitted from all the components shown in the embodiments, or components of different types may be appropriately combined.

[0054] Although specific examples are described in this disclosure, they are intended in all respects to be illustrative and not limiting (for understanding the technology of the present disclosure). Those skilled in the art will recognize that there are numerous combinations of hardware, software, and firmware suitable for implementing the technology of the present disclosure. For example, the software described can be implemented in a wide variety of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, Java (registered trademark), etc.

[0055] Furthermore, in the above-described embodiment, the control lines and information lines are those that are considered necessary for the explanation, and not all control lines and information lines in the product are necessarily shown. All components may be interconnected.

[0056] In addition, other implementations of the present disclosure will be apparent to those skilled in the art from consideration of the present embodiments. The specification and examples are exemplary only, with the scope and spirit of the technology of the present disclosure being indicated by the following claims. [Explanation of symbols]

[0057] 1 Daily report analysis device 11 Input Devices 12 Output Devices 13 Central processing unit 14 Main memory 15 Auxiliary storage 51 Pretreatment 52 Evaluation value calculation process 53 Deviation calculation process 54 Confirmation priority calculation process 91 Description Judgment Processing 92 Calculation of updated evaluation value 93 Weighted Average Processing 111 Feature Extraction Processing 151 Daily Report Management Table 152 Evaluation Index Setting Table 153 Progress Settings Table

Claims

1. A daily report analysis device that outputs priorities of daily reports to be confirmed from daily reports of multiple projects, a storage device that stores a daily report analysis program for determining the priority of the daily reports of the plurality of cases and outputting information on the priority; a processor that reads the daily report analysis program from the storage device and executes it; The processor: A preprocessing step of converting the text information of the daily report into a text vector represented by a numerical vector; an evaluation value calculation process for calculating an evaluation value for each preset evaluation index using the text vector; a deviation calculation process for calculating a progress deviation based on the evaluation value of each evaluation index and the progress registration value reported in the daily report; a confirmation priority calculation process that calculates a confirmation priority indicating a confirmation priority order for each of the projects based on the progress deviation degrees of the multiple projects obtained by the deviation degree calculation process; an output process for visualizing the plurality of cases in a line based on the numerical values ​​of the confirmation priorities; A daily report analysis device that performs the following.

2. In claim 1, The processor calculates the evaluation value for each evaluation index using machine learning or deep learning.

3. In claim 1, The processor further comprises: A process of calculating updated evaluation values, which are evaluation values ​​of each of the evaluation indexes, for the daily report to be analyzed using the text vectors; A process of calculating a description probability p of each evaluation index in the daily report to be analyzed using the text vector; A daily report analysis device that generates and outputs the evaluation value by using the description probability p to calculate a weighted average of the pre-update evaluation value, which is the evaluation value in a daily report for the same project that was registered before the daily report to be analyzed was registered, and the post-update evaluation value.

4. In claim 3, The processor generates the evaluation value by calculating the weighted average as follows: the description probability p x the post-update evaluation value + (1 - the description probability p) x the pre-update evaluation value.

5. In claim 1, The processor: In the confirmation priority calculation process, the confirmation priority is calculated using paired information of a given item importance and the progress deviation degree for the plurality of items; In the output process, the daily report analysis device arranges and visualizes the plurality of cases based on the confirmation priority on a two-dimensional plane formed by two axes, the case importance and the progress deviation.

6. In claim 1, The processor further executes a process of extracting a feature vector of the evaluation value using the text vector; the processor calculates the evaluation value for each of the evaluation indexes by applying machine learning or deep learning to the feature vector in the evaluation value calculation process; The machine learning or deep learning applied to the feature vectors is trained so that, for two different text vectors, if the training data for the evaluation values ​​of the text vectors registered in advance is the same, the similarity between the two feature vectors is high, and if the training data for the evaluation values ​​is different, the similarity between the two feature vectors is low.

7. A daily report analysis method for outputting priorities of daily reports to be confirmed from daily reports of multiple projects, a daily report analysis step in which a processor reads a daily report analysis program from a storage device that stores the daily report analysis program to determine priorities of daily reports of the plurality of cases and output information on the priorities, and executes the daily report analysis program; The daily report analysis step includes: The processor converts the text information of the daily report into a text vector represented by a numerical vector; The processor calculates an evaluation value for each of the predetermined evaluation indicators using the text vector; the processor calculates a progress deviation degree based on the evaluation value of each evaluation index and the progress registration value reported in the daily report; the processor calculates a confirmation priority indicating a confirmation priority for each of the plurality of cases based on a progress deviation degree of the plurality of cases; the processor arranges and visualizes the plurality of cases based on the numerical values ​​of the confirmation priorities; Daily report analysis methods, including:

8. A daily report analysis program for causing a computer to execute a daily report analysis process for outputting priorities of daily reports to be confirmed from daily reports of a plurality of cases, the daily report analysis program is a program for determining the priority of the daily reports of the plurality of cases and outputting information on the priority; The daily report analysis program is installed on the computer. A process of converting the text information of the daily report into a text vector represented by a numerical vector; A process of calculating an evaluation value for each of the predetermined evaluation indexes using the text vector; a process of calculating a progress deviation degree based on the evaluation value of each evaluation index and the progress registration value reported in the daily report; A process of calculating a confirmation priority indicating a confirmation priority for each of the plurality of projects based on the progress deviation of the plurality of projects; A process of arranging and visualizing the plurality of cases based on the numerical values ​​of the confirmation priorities; A daily report analysis program that executes the following.

Images