Information processing device, determination method, and determination program

Abstract

To enable management of work time for each project.SOLUTION: An information processing device is provided, comprising an acquisition unit for acquiring a file accessed by a worker, and a determination unit configured to determine a project involving work that uses the file on the basis of text data included in the file.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an information processing device, a determination method, and a determination program. [Background technology]

[0002] As one technique for managing the man-hours of work using a computer, the following man-hours management device has been proposed (see, for example, Patent Document 1). For example, the man-hours management device calculates the man-hours of a group of files by excluding, from a first period during which an object corresponding to the group of files is active, a second period between two consecutive operation times, the second period having a length equal to or greater than a threshold value. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2012-164126 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the above-mentioned man-hour management device merely calculates the work time for a file, and therefore has an aspect that makes it difficult to manage work time on a project-by-project basis.

[0005] An object of the present invention is to realize management of working hours on a project-by-project basis. [Means for solving the problem]

[0006] An information processing device according to one aspect of the present invention has an acquisition unit that acquires a file accessed by a worker, and a determination unit that determines to which project a task using the file belongs based on text data contained in the file.

[0007] In a determination method according to one aspect of the present invention, a computer executes a process of obtaining a file accessed by a worker, and determining to which project the work using the file belongs based on text data contained in the file.

[0008] A determination program according to one aspect of the present invention causes a computer to execute a process of acquiring a file accessed by a worker and determining the project to which a task using the file belongs based on text data contained in the file. Effect of the Invention

[0009] According to one embodiment, it is possible to realize management of working hours on a project-by-project basis. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a man-hour management system. [Diagram 2] FIG. 2 is a block diagram illustrating an example of a functional configuration of the server device. [Diagram 3] FIG. 3 is a schematic diagram illustrating an example of a machine learning model. [Figure 4] FIG. 4 is a diagram illustrating an example of man-hour management data. [Diagram 5] FIG. 5 is a diagram showing an example of the counting result. [Figure 6] FIG. 6 is a diagram showing an example of a display on the administrator terminal. [Figure 7] FIG. 7 is a diagram showing an example of the counting result. [Figure 8] FIG. 8 is a diagram showing an example of a display on the administrator terminal. [Figure 9] FIG. 9 is a flowchart showing the procedure of the training process. [Figure 10] FIG. 10 is a flowchart showing the procedure of the determination process. [Figure 11] FIG. 11 is a schematic diagram illustrating an example of a machine learning model. [Figure 12]FIG. 12 is a diagram illustrating an example of a hardware configuration. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the information processing device, the determination method, and the determination program according to the present application (hereinafter, referred to as "embodiments") will be described with reference to the attached drawings. Each embodiment is merely an example, and the configuration, structure, functions, numerical values ​​used to specify these, and usage scenes are not limited by such examples. Each embodiment can be adaptively combined to the extent that the processing contents are not contradictory.

[0012] <Overall composition> Fig. 1 is a diagram showing an example of the configuration of a man-hour management system. The man-hour management system 1 shown in Fig. 1 provides a man-hour management function for managing the work time during which a worker performs work using a computer, that is, the so-called man-hours.

[0013] The "work" referred to here may include, by way of example only, work related to a project to design or develop a product, system, service, etc.

[0014] 1, the man-hour management system 1 may include a server device 10, worker terminals 30A to 30N, and an administrator terminal 50. Hereinafter, when there is no need to distinguish between the individual worker terminals 30A to 30N, the worker terminals 30A to 30N may be referred to as "worker terminal 30."

[0015] The server device 10, the worker terminal 30, and the administrator terminal 50 may be communicably connected to each other via any network NW. The network NW may be realized by any technology, such as the Internet technology, industrial communication standards, or low-power wireless communication standards for the Internet of Things (IoT), regardless of whether it is wired or wireless.

[0016] The server device 10 is an example of an information processing device that provides the above-mentioned man-hour management function. For example, the server device 10 can provide a man-hour management service corresponding to the above-mentioned man-hour management function as a cloud service by being realized as a PaaS (Platform as a Service) type or SaaS (Software as a Service) type application. In addition, the server device 10 may be realized as a server that provides the above-mentioned man-hour management function on-premise.

[0017] The worker terminal 30 is an example of a computer used by a worker for the above-mentioned work. For example, the worker terminal 30 may be realized by any computer, such as a personal computer, a smartphone, a tablet terminal, or a wearable terminal.

[0018] The manager terminal 50 is an example of a computer used by a manager who manages man-hours in the project. Like the worker terminal 30, the manager terminal 50 may be realized by any computer, such as a personal computer, a smartphone, a tablet terminal, or a wearable terminal.

[0019] <One aspect of the issue> As explained in the Background section above, conventional technologies such as the man-hour management device described in Patent Document 1 merely calculate the work time for a file, making it difficult to manage work time on a project-by-project basis.

[0020] In order to realize project-based work time management in such conventional technology, manual intervention is required, such as manually linking files used in a work to a project or imposing restrictions on the naming rules of files used in a work, such as giving file names that can identify a project.

[0021] <One aspect of the problem-solving approach> Therefore, in this embodiment, as part of the man-hour management function, a determination function is provided that determines the project to which the work using a file belongs based on the text data contained in the file accessed by the worker terminal 30.

[0022] This determination function distinguishes the work performed using the worker terminal 30 by project, so that the files used in the work can be linked to the project without human intervention. Therefore, the above determination function makes it possible to manage the work time by project, i.e., to manage man-hours. Furthermore, the above determination function can be realized without necessarily using the calculation of a complex algorithm, so that the calculation cost related to the above determination function can also be reduced.

[0023] <Configuration of Server Device 10> Next, a functional configuration example of the server device 10 having a man-hour management function in which the above-mentioned judgment function is packaged will be described. Fig. 2 is a block diagram showing a functional configuration example of the server device 10. Fig. 2 shows a schematic diagram of blocks related to the man-hour management function of the server device 10.

[0024] As shown in Fig. 2, the server device 10 has a communication control unit 11, a storage unit 13, and a control unit 15. Note that Fig. 2 merely illustrates a selection of functional units related to the man-hour management function described above, and the server device 10 may include functional units other than those illustrated.

[0025] The communication control unit 11 is a functional unit that controls communication with other devices such as the worker terminal 30 and the manager terminal 50. As just one example, the communication control unit 11 can be realized by a network interface card. As one aspect, the communication control unit 11 can receive an operation log of the worker terminal 30 from the worker terminal 30. As another aspect, the communication control unit 11 can receive a request for uploading training data used for training a machine learning model for project judgment described later, training a machine learning model, a request for outputting man-hours, and the like from the manager terminal 50. In addition, the communication control unit 11 can output various data related to man-hours to the manager terminal 50.

[0026] The storage unit 13 is a functional unit that stores various data. As an example, the storage unit 13 is realized by an internal, external, or auxiliary storage of the server device 10. For example, the storage unit 13 stores work log data 13A, training data 13B, a machine learning model 13C, and man-hour management data 13D. The work log data 13A, the training data 13B, the machine learning model 13C, and the man-hour management data 13D will be described together when reference, generation, or registration is performed.

[0027] The control unit 15 is a functional unit that performs overall control of the server device 10. For example, the control unit 15 can be realized by a hardware processor. As shown in Fig. 2, the control unit 15 has a calculation unit 15A, an acquisition unit 15B, a determination unit 15C, a training unit 15D, an update unit 15E, and an output unit 15F. The control unit 15 may be realized by hardwired logic or the like.

[0028] The calculation unit 15A is a processing unit that calculates the work time for a file based on the period of access to the file by the worker terminal 30. As merely one example, the calculation unit 15A can calculate the work time for a file according to the technique described in the above-mentioned Patent Document 1.

[0029] More specifically, the calculation unit 15A can execute the following processes in parallel for each worker terminal 30. For example, the calculation unit 15A collects operation logs from the worker terminal 30. When the operation log collected in this way is a file close event, the calculation unit 15A calculates the active period of the file. The "active period" here refers to a period during which an object, such as a window, generated by an application executing a file is active. In addition, the "file" does not necessarily have to be one saved in the worker terminal 30. For example, the file may be a file stored in an external device such as a file server, or may be a file stored in an auxiliary storage device such as a removable medium.

[0030] Then, the calculation unit 15A calculates the work time for the file by subtracting the no-operation period from the active time. For example, the calculation unit 15A extracts operation logs whose operation times are included in the active period from among the operation logs collected from the worker terminal 30. Then, the calculation unit 15A detects a no-operation period in which the interval between the operation times of two consecutive operation logs is equal to or greater than a threshold, and calculates the work time for the file by subtracting the no-operation period from the active time.

[0031] Thereafter, the calculation unit 15A generates a work log including the identification information of the worker currently logged in to the worker terminal 30, the file accessed by the worker terminal 30, and the work time for the file, and then adds and registers the work log data 13A stored in the storage unit 13. Hereinafter, a file used in work by the worker terminal 30 may be referred to as a "work file".

[0032] In this way, a set of work logs with a division from an open event to a close event of a work file as a work unit is accumulated as the work log data 13A. Note that, here, as an example of the work time for a file, an example of calculating a period during which an object generated from a file is active and not unoperated is given, but this is not limiting. For example, the calculation unit 15A calculates the above-mentioned active period as the work time for a file. In addition, the calculation unit 15A can also calculate, as the work time for a file, an access period corresponding to the period from the operation time when an open event of the work file occurs to the operation time when a close event occurs.

[0033] The acquisition unit 15B is a processing unit that acquires a work log. In this way, the trigger for starting the acquisition of the work log may be any timing. As just one example, the acquisition unit 15B can acquire a work log at a specific time, such as 18:00 or 24:00, at a fixed cycle, for example, a cycle of one day. For example, the acquisition unit 15B can acquire one work log from a set of N work logs related to each M workers. Then, the acquisition unit 15B extracts a work file included in the n-th work log related to the m-th worker. As another example, the acquisition unit 15B can acquire a new work log every time the new work log is added to the work log data 13A.

[0034] The determination unit 15C is a processing unit that determines the project to which the work using the work file belongs based on the text data included in the work file. The "text data" referred to here is not limited to the text included in the file content. For example, the text data may include file metadata in addition to the file name.

[0035] Such project determination can be realized, as an example only, by a machine learning model 13C that executes a class classification task that inputs text data and outputs confidence levels for each class of project. For example, the machine learning model 13C may be realized by a recurrent neural network used in natural language processing, such as a Long Short-Term Memory (LSTM), or a neural network such as a Transformer. The machine learning model 13C is not limited to such a neural network, and may be realized by any other machine learning model, such as a support vector machine or boosting.

[0036] The training unit 15D is a processing unit that trains the machine learning model 13C that the judgment unit 15C uses for project judgment.

[0037] FIG. 3 is a schematic diagram for explaining an example of a machine learning model. As shown in FIG. 3, a machine learning model 13C is used for project determination. A training dataset 13B including training data in which word strings and correct answer labels of projects are associated with each other can be used for training this machine learning model 13C. Such a training dataset 13B may be uploaded via an administrator terminal 50. For example, FIG. 3 illustrates, as merely one example of the training dataset 13B, three pieces of training data in which word strings are associated with the correct answer labels of three classes related to projects, "Project X", "Project Y", and "Project Z".

[0038] As one aspect, in the training phase, the feature vector in which the word sequence is quantified can be used as an explanatory variable of the machine learning model 13C, and the correct answer label can be used as the objective variable of the machine learning model 13C, and the machine learning model 13C can be trained according to any machine learning algorithm, for example, deep learning.

[0039] For example, the parameters of the machine learning model 13C are updated by backpropagating the output of the machine learning model 13C to which the feature vector of the word string is input and the loss between the correct label to the machine learning model 13C.

[0040] Such parameter updates are repeated until an arbitrary termination condition is met. For example, examples of the termination condition include a condition that a specified number of epochs is reached, or a condition that the amount of parameter updates is less than a threshold. As a result, a trained machine learning model 13C is generated.

[0041] As another aspect, in the inference phase, the determination unit 15C digitizes the text data included in the work file acquired by the acquisition unit 15B. For example, the determination unit 15C executes natural language processing, such as morphological analysis and named entity extraction, on the text data included in the work file. As a result, a word string is obtained from the text data of the work file. Then, the determination unit 15C converts the word string obtained from the text data of the work file into numerical data, such as a feature vector. At this time, the determination unit 15C can narrow down the words included in the word string to some words, such as nouns and proper nouns, and perform digitization. For example, as shown in FIG. 3, the determination unit 15C can narrow down the words corresponding to the labels of specific named entities such as organizations, place names, and product names, such as product name A, part name D, and work name G, and perform digitization.

[0042] As just one example, the digitization of a word string may be achieved according to a feature extraction method such as Bag of Words (BoW) or Term Frequency (TF)-Inverse Document Frequency (IDF).

[0043] When BoW or TF-IDF is applied, a set of words that appear in the training dataset 13B is treated as a vocabulary, and a feature vector is obtained by calculating the number of occurrences, occurrence frequency, and inverse document frequency of each word contained in the vocabulary in the input word string.

[0044] Then, the determination unit 15C inputs the feature vector of the word string to the trained machine learning model 13C. The machine learning model 13C to which the feature vector of the word string has been input in this manner outputs the confidence level for each class of the project.

[0045] For example, in the example shown in FIG. 3, "80%" is output as the confidence level for class "Project X." Furthermore, "15%" is output as the confidence level for class "Project Y." Furthermore, "5%" is output as the confidence level for class "Project Z." In this case, just as an example, the class "Project X" with the highest confidence level can be determined as the project determination result.

[0046] Returning to the explanation of FIG. 2, the update unit 15E is a processing unit that updates the man-hour management data 13D stored in the storage unit 13.

[0047] Fig. 4 is a diagram showing an example of man-hour management data 13D. As shown in Fig. 4, man-hour management data 13D may be data in which items such as date, worker, project, and working time (h) are associated with each other. In this way, according to man-hour management data 13D, the working time of each worker, that is, the so-called man-hours, is managed in a specific unit time, for example, a bucket size of one day.

[0048] As an example only, when the project is judged by the judgment unit 15C, the update unit 15E executes the following process. That is, the update unit 15E updates the identification information of the mth worker whose work log is acquired by the acquisition unit 15B and the work time of the data entry corresponding to the nth work project whose judgment is executed by the judgment unit 15C, in the man-hour management data 13D. For example, the update unit 15E adds the work time included in the work log acquired by the acquisition unit 15B to the work time stored in the data entry. By updating the work time in this way, a cumulative sum of the work times, that is, a so-called total, is obtained for each worker.

[0049] The output unit 15F is a processing unit that outputs various information to the manager terminal 50. As one aspect, the output unit 15F can display various data related to man-hours on the manager terminal 50. Such display control can also be executed when a request for outputting man-hours is received from the manager terminal 50. In addition, the above-mentioned display control can be executed as a report or an alert when a specific condition is satisfied. For example, a report can be output daily, weekly, monthly, or yearly, and an alert can be output when the working hours exceed a threshold value.

[0050] As merely one example, the output unit 15F can calculate and display statistical values, such as aggregate values, average values, and median values, of the work hours included in the man-hour management data 13D for each project. At this time, the output unit 15F can aggregate the work hours of data entries included in a specified time range in the man-hour management data 13D by a specified bucket size, such as daily, weekly, monthly, quarterly, half-yearly, or annually. Note that the time range and bucket size may be specified by a user definition by an administrator or the like, or by system definition.

[0051] Fig. 5 is a diagram showing an example of the aggregation result. Fig. 5 shows an aggregation result 20 in which data for a three-month time range of months D, E, and F in the man-hour management data 13D shown in Fig. 4 is aggregated monthly. Fig. 5 shows three projects, project X, project Y, and project Z, as examples of projects, and also shows three people, Mr. K, Mr. L, and Mr. M, as examples of workers.

[0052] As shown in Figure 5, the tabulation result 20X includes the work hours of Project X in month D, month E, and month F for two workers, Mr. K and Mr. L. Note that in Figure 5, since the tabulation value of Mr. M's work hours of Project X is "0" in all months D, E, and F, the display of the work hours related to Mr. M is omitted.

[0053] Furthermore, the summary result 20Y includes the work hours for project Y in month D, the work hours for project Y in month E, and the work hours for project X in month Y for three workers, Mr. K, Mr. L, and Mr. M.

[0054] Furthermore, in the tabulation result 20Z, the working hours of Project Z in month D, month E, and month Y are tabulated for three workers, Mr. K, Mr. L, and Mr. M. In addition, in Fig. 5, since the tabulated value of Mr. K's working hours of Project X is "0" in all months D, E, and F, the display of the working hours related to Mr. K is omitted.

[0055] When such a project-based work time compilation result 20 is displayed and output, the project-based work time compilation result 20 can be graphed as shown in FIG. 6, which is merely one example. FIG. 6 is a diagram showing a display example of the manager terminal 50. FIG. 6 shows an example of a graph in which the work time compilation result 20 shown in FIG. 5 is made into a bar graph. As shown in FIG. 6, the compilation value of the work time of each worker, K, L, and M, is displayed as a bar graph for each of the three projects, project X, project Y, and project Z. This type of display makes it easy to grasp the individual man-hours of K, L, and M on a project-by-project basis, and also enables a relative evaluation of the man-hours of the three workers.

[0056] Fig. 7 is a diagram showing an example of the aggregation result. Fig. 7 shows an aggregation result 40 in which data for a three-month time range of months D, E, and F in the man-hour management data 13D shown in Fig. 4 is aggregated monthly. Note that Fig. 7 also shows three projects, project X, project Y, and project Z, as examples of projects, and excerpts of three people, Mr. K, Mr. L, and Mr. M, as examples of workers.

[0057] As shown in Fig. 7, the tabulation result 40K includes K's work hours in month D, month E, and month F for projects X and Y. Note that in Fig. 7, since the tabulation value of K's work hours for project Z is "0" in all months D, E, and F, the display of the work hours for project Z is omitted.

[0058] Furthermore, in the summary result 40L, Mr. L's working hours in month D, month E, and month Y are summarized for each of project X, project Y, and project Z.

[0059] Furthermore, in the tabulation result 40M, M's working hours in month D, month E, and month Y are tabulated for each of project Y and project Z. Note that in Fig. 7, the tabulated value of M's working hours for project X is "0" in all of months D, E, and F, so the display of the working hours for project X is omitted.

[0060] Such tabulation results 40 of work times by worker can also be graphed. FIG. 8 is a diagram showing an example of display on the manager terminal 50. FIG. 8 shows, as an example of graphing, an example in which tabulation results 40L of L's work times among tabulation results 40 of work times by worker shown in FIG. 7 are graphed as a pie chart. As shown in FIG. 8, the tabulation values ​​of L's work times are displayed as a pie chart 60 for three projects, project X, project Y, and project Z. Displaying such a pie chart 60 makes it easy to grasp the breakdown of the man-hours of each project out of L's total work.

[0061] In addition, in Figures 5 to 8, aggregated values ​​are given as examples of statistical values ​​of working time, but other statistical values, such as average values ​​and median values, can also be calculated and displayed for time ranges such as days, weeks, months, and years.

[0062] 6 and 8, display output is given as an example of an output form by the output unit 15F, but print output, audio output, etc. may be executed by the output unit 15F. Also, here, the administrator terminal 50 is given as an example of an output destination, but the output destination is not limited to a user terminal. For example, the output unit 15F can output numerical values ​​and chart data related to man-hours as a data file in a format executable by any application, such as a spreadsheet software. In addition, the output unit 15F can output numerical values ​​and chart data related to man-hours to a back-end function of the above-mentioned judgment function, such as a service that performs analysis and analysis related to man-hour management.

[0063] <Processing flow> Next, a flow of processing of the server device 10 according to the present embodiment will be described. Here, (1) training processing executed by the server device 10 will be described first, followed by (2) judgment processing.

[0064] (1) Training process 9 is a flowchart showing the procedure of the training process, which is merely an example and can be started when a request for training the machine learning model 13C is received.

[0065] 9, the training unit 15D executes a loop process 1 that repeats the processes of the following step S101 and the following step S102 until an arbitrary termination condition is satisfied. For example, the termination condition may be a condition that a specified number of epochs is reached, or a condition that the update amount of a parameter is less than a threshold value.

[0066] Furthermore, the training unit 15D executes a loop process 2 that repeats the processes of step S101 and step S102 described below a number of times corresponding to the number K of training data included in the training data set 13B.

[0067] That is, the training unit 15D inputs a feature vector obtained by digitizing a word string included in the k-th training data to the machine learning model 13C (step S101). Next, the training unit 15D updates the parameters of the machine learning model 13C based on the loss between the class label of the project output by the machine learning model 13C to which the feature vector was input in step S101 and the correct label included in the k-th training data (step S102).

[0068] By repeating such loop process 2, one epoch of training on the training data set 13B is completed. Furthermore, by repeating loop process 1, the parameters of the machine learning model 13C converge. As a result, the trained machine learning model 13C is generated.

[0069] Thereafter, the training unit 15D stores the trained machine learning model 13C in the storage unit 13 (step S103), and ends the process.

[0070] (2) Judgment process 10 is a flowchart showing the procedure of the training process. This process can be executed periodically, for example, at a specific time every day, for example, at 6 p.m. or midnight.

[0071] As shown in FIG. 10, the acquisition unit 15B executes a loop process 1 that repeats the processes from step S301 to step S304 described below a number of times corresponding to the number M of workers.

[0072] Furthermore, the acquisition unit 15B executes loop processing 2, which repeats the processes from step S301 to step S304 described below a number of times corresponding to the number N of work logs related to the mth worker among the work logs included in the work log data 13A.

[0073] That is, the acquiring unit 15B acquires the n-th work log related to the m-th worker from the work log data 13A (step S301). Next, the acquiring unit 15B extracts the work file included in the work log acquired in step S301 (step S302).

[0074] Next, the determination unit 15C determines the project to which the n-th work log belongs based on the project class label output by the machine learning model 13C to which the word string contained in the work file extracted in step S302 is input (step S303).

[0075] Then, the update unit 15E updates the work time of the data entry corresponding to the project to which the m-th worker's identification information and the n-th work log belong, in the man-hour management data 13D (step S304).

[0076] By repeating this loop process 2, the work times included in the N work logs related to the m-th worker are summed up in a state of being classified by project. Furthermore, by repeating loop process 1, the total value of the daily work times for M workers is obtained by project.

[0077] <One aspect of the effect> As described above, the server device 10 according to the present embodiment determines the project to which the work using the file belongs based on the text data included in the file accessed by the worker terminal 30. This allows the work using the worker terminal 30 to be distinguished by project, so that the file used in the work can be linked to the project without human intervention. Therefore, the above-mentioned determination function makes it possible to manage the work time by project, i.e., to manage man-hours.

[0078] <Other embodiments> Although the embodiment of the present invention has been described above, the present invention is applicable to various applications, and may be embodied in various different forms other than the above-described embodiment.

[0079] In the above embodiment, an example of determining a project to which an operation using a work file belongs is given, but the present invention is not limited to this. For example, the determination unit 15C can further determine not only the project but also the process to which the operation using the work file belongs among the multiple processes into which the project is divided. Such determination of the project and process can be realized, as an example only, by having the machine learning model 13C output a class label of "project name + process name".

[0080] FIG. 10 is a schematic diagram for explaining an example of a machine learning model. As shown in FIG. 10, a machine learning model 13C is also used to determine projects and processes. A training data set 13B including training data in which a word string is associated with a correct answer label of a class including a project name and a process name can be used for training this machine learning model 13C. For example, FIG. 10 excerpts four training data in which a word string is associated with each correct answer label of classes "Project X1", "Project X2", "Project X3", and "Project Z3" as only one example of the training data set 13B.

[0081] Here, the label of the class "Project X1" includes the project name "Project X" + process name "Process 1". Furthermore, the label of the class "Project X2" includes the project name "Project X" + process name "Process 2". Furthermore, the label of the class "Project X3" includes the project name "Project X" + process name "Process 3". Furthermore, the label of the class "Project Z3" includes the project name "Project Z" + process name "Process 3".

[0082] As one aspect, in the training phase, the feature vector in which the word sequence is quantified can be used as an explanatory variable of the machine learning model 13C, the correct answer label can be used as the objective variable of the machine learning model 13C, and the machine learning model 13C can be trained according to any machine learning algorithm.

[0083] For example, the parameters of the machine learning model 13C are updated by backpropagating the output of the machine learning model 13C to which the feature vector of the word string is input and the loss between the correct label to the machine learning model 13C.

[0084] Such parameter updates are repeated until an arbitrary termination condition is met. For example, examples of the termination condition include a condition that a specified number of epochs is reached, or a condition that the amount of parameter updates is less than a threshold. As a result, a trained machine learning model 13C is generated.

[0085] As another aspect, in the inference phase, the determination unit 15C digitizes the text data included in the work file acquired by the acquisition unit 15B. Then, the determination unit 15C inputs the feature vector of the word string to the trained machine learning model 13C. The machine learning model 13C to which the feature vector of the word string is input in this manner outputs the confidence level by class including the project name and the process name.

[0086] For example, in the example shown in FIG. 10, "5%" is output as the confidence level for class "Project X1". Furthermore, "70%" is output as the confidence level for class "Project X2". Furthermore, "15%" is output as the confidence level for class "Project X3". Furthermore, "0%" is output as the confidence level for class "Project Z3". In this case, just as an example, since the confidence level for class "Project X2" is the highest, the project name can be determined to be "Project X" and the process name can be determined to be "Process 2".

[0087] <Numbers, etc.> The matters described in the above embodiment, such as specific examples of the training method and inference method of the machine learning model, are merely examples and may be changed. In addition, the order of the processes in the flowcharts described in the embodiment may be changed within a range that does not cause inconsistencies.

[0088] <System> The information including the processing procedures, control procedures, specific names, various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, any one or more of the functional units among the calculation unit 15A, the acquisition unit 15B, the judgment unit 15C, the training unit 15D, the update unit 15E and the output unit 15F may be configured as separate devices.

[0089] In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure. In other words, all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc. Each configuration may be a physical configuration.

[0090] For example, Fig. 1 shows an example in which the above-mentioned man-hour management function is provided by a client-server system, but this is merely an example, and the above-mentioned man-hour management function does not necessarily have to be provided by a client-server system. For example, the above-mentioned man-hour management function may be provided by being executed on a worker terminal 30. In this case, the worker terminal 30 holds man-hour management data that manages the work time of a worker using the worker terminal 30. For this reason, the above-mentioned man-hour management function can be provided by sharing man-hour management data by P2P, or by sharing man-hour management data with one worker terminal 30 as a master and the remaining worker terminals 30 as slaves.

[0091] Furthermore, each processing function performed by each device may be realized, in whole or in part, by a CPU (Central Processing Unit) and a program analyzed and executed by the CPU, or may be realized as hardware using wired logic.

[0092] <Hardware configuration> Moreover, the various processes described in the above embodiments can be realized by executing a prepared program on a computer such as a personal computer or a workstation. Therefore, an example of a computer that executes a determination program having the same functions as those in the first and second embodiments will be described below with reference to FIG.

[0093] Fig. 12 is a diagram showing an example of a hardware configuration. As shown in Fig. 12, the computer 100 has an operation unit 110a, a speaker 110b, a camera 110c, a display 120, and a communication unit 130. Furthermore, the computer 100 has a CPU 150, a ROM 160, a HDD 170, and a RAM 180. These units 110 to 180 are connected via a bus 140.

[0094] As shown in Fig. 12, the HDD 170 stores a judgment program 170a that performs the same functions as the calculation unit 15A, the acquisition unit 15B, the judgment unit 15C, the training unit 15D, the update unit 15E, and the output unit 15F shown in the above-mentioned first embodiment. This judgment program 170a may be integrated or separated like the components of the calculation unit 15A, the acquisition unit 15B, the judgment unit 15C, the training unit 15D, the update unit 15E, and the output unit 15F shown in Fig. 2. That is, the HDD 170 does not necessarily have to store all the data shown in the above-mentioned first embodiment, and it is sufficient that the data used for processing is stored in the HDD 170.

[0095] Under such an environment, the CPU 150 reads out the judgment program 170a from the HDD 170 and loads it in the RAM 180. As a result, the judgment program 170a functions as a judgment process 180a as shown in FIG. 12. The judgment process 180a loads various data read out from the HDD 170 in an area of ​​the storage area of ​​the RAM 180 that is assigned to the judgment process 180a, and executes various processes using the loaded various data. For example, the processes shown in FIG. 9 and FIG. 10 may be included as examples of the processes executed by the judgment process 180a. Note that in the CPU 150, all of the processing units shown in the above-mentioned first embodiment do not necessarily need to operate, and it is sufficient that the processing units corresponding to the processes to be executed are virtually realized.

[0096] The above-mentioned determination program 170a does not necessarily have to be stored in the HDD 170 or the ROM 160 from the beginning. For example, the determination program 170a is stored in a "portable physical medium" such as a flexible disk, a so-called FD, a CD-ROM, a DVD disk, a magneto-optical disk, or an IC card that is inserted into the computer 100. The computer 100 may then acquire and execute the determination program 170a from the portable physical medium. The determination program 170a may also be stored in another computer or a server device connected to the computer 100 via a public line, the Internet, a LAN, a WAN, or the like. The thus-stored determination program 170a may be downloaded to the computer 100 and then executed.

[0097] <Other> Some examples of combinations of the disclosed technical features are set out below.

[0098] (1) an acquisition unit for acquiring a file accessed by an operator; a determination unit that determines a project to which an operation using the file belongs based on text data included in the file; 13. An information processing device comprising:

[0099] (2) The information processing device described in (1) is characterized in that the determination unit inputs a word string extracted from the text data to a machine learning model that takes a word string as input and outputs a class of project name, and determines the project to which the work using the file belongs based on the class output by the machine learning model.

[0100] (3) The information processing device described in (2), further comprising a training unit that trains parameters of the machine learning model by performing machine learning with the feature vector obtained from the word string as an explanatory variable of the machine learning model and the correct label of the project name class as the objective variable of the machine learning model.

[0101] (4) The information processing device described in (1) is characterized in that the determination unit inputs a word string extracted from the text data to a machine learning model that takes a word string as input and outputs a class including a project name and a process name, thereby determining the project to which the work using the file belongs based on the class output by the machine learning model, and determining the process to which the work using the file belongs among multiple processes included in the project.

[0102] (5) The information processing device described in (4), further comprising a training unit that trains parameters of the machine learning model by performing machine learning using a feature vector obtained from the word string as an explanatory variable of the machine learning model and a correct answer label of a class including the project name and the process name as a target variable of the machine learning model.

[0103] (6) The information processing device according to any one of (1) to (5), further comprising an output unit that links the project determined by the determination unit with a working time calculated from the access period of the file and outputs the linked project.

[0104] (7) The information processing device according to (6), wherein the output unit outputs a compilation result in which the working hours are compiled for each project.

[0105] (8) The information processing device according to (7), wherein the output unit outputs a bar graph or a pie chart generated from the tallying results.

[0106] (9) Obtain the files that were accessed by the worker, determining a project to which an operation using the file belongs based on text data included in the file; A determination method characterized in that the processing is executed by a computer.

[0107] (10) Obtain the files that were accessed by the worker, determining a project to which an operation using the file belongs based on text data included in the file; A determination program that causes a computer to execute a process. [Explanation of symbols]

[0108] 1. Man-hour management system 10. Server equipment 11 Communication control section 13 Storage section 13A Work log data 13B Training Data 13C Machine Learning Model 13D Man-hour management data 15 Control section 15A calculation section 15B Acquisition Department 15C Judgment section 15D Training Department 15E Update section 15F Output section 30 Worker terminal

Claims

1. an acquisition unit for acquiring a file accessed by an operator; a determination unit that determines a project to which an operation using the file belongs based on text data included in the file; 13. An information processing device comprising:

2. The information processing device according to claim 1, characterized in that the determination unit inputs a word string extracted from the text data to a machine learning model that takes a word string as input and outputs a class of project names, and determines the project to which the work using the file belongs based on the class output by the machine learning model.

3. The information processing device according to claim 2, further comprising a training unit that trains parameters of the machine learning model by performing machine learning with a feature vector obtained from the word string as an explanatory variable of the machine learning model and a correct answer label of the project name class as a target variable of the machine learning model.

4. The information processing device according to claim 1, characterized in that the determination unit inputs a word string extracted from the text data to a machine learning model that takes a word string as input and outputs a class including a project name and a process name, thereby determining a project to which the work using the file belongs based on the class output by the machine learning model, and determining the process to which the work using the file belongs among a plurality of processes included in the project.

5. The information processing device according to claim 4, further comprising a training unit that trains parameters of the machine learning model by performing machine learning with a feature vector obtained from the word string as an explanatory variable of the machine learning model and a correct answer label of a class including the project name and the process name as a target variable of the machine learning model.

6. 2 . The information processing apparatus according to claim 1 , further comprising an output unit that outputs the project determined by the determination unit in association with a working time calculated from an access period of the file.

7. The information processing apparatus according to claim 6 , wherein the output unit outputs a result of tallying up the work time for each project.

8. The information processing apparatus according to claim 7 , wherein the output unit outputs a bar graph or a pie chart generated from the counting result.

9. Obtain the files that were accessed by the worker, determining a project to which an operation using the file belongs based on text data included in the file; A determination method characterized in that the processing is executed by a computer.

10. Obtain the files that were accessed by the worker, determining a project to which an operation using the file belongs based on text data included in the file; A determination program that causes a computer to execute a process.

Images