Estimation program, estimation system, and estimation method

Abstract

To provide a technique for providing new information based on presentation materials.SOLUTION: An estimation program causing a computer to function as an acquisition unit, an estimation unit, and an output unit, where the acquisition unit acquires material information indicating information included in a presentation material used for an enterprise activity, the estimation unit performs input to an estimation model based on the material information, and acquires an estimation result regarding a result of the enterprise activity as output of the estimation model, and the output unit performs output based on the estimation result.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an estimation program, an estimation system, and an estimation method. [Background technology]

[0002] In business activities, presentations are routinely given for various purposes, and techniques for evaluating presentation materials used in such situations are known. For example, Patent Document 1 describes a technique for predicting audience responses using a model based on the characteristics of a group of slides and the attributes of the audience, and selecting the slides to be used based on the results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-110516 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 merely predicted the reactions of the audience and assisted them in selecting slides from a set of slides prepared in advance, and was therefore unable to assist in the creation of presentation materials or predict the effectiveness of the presentation.

[0005] In view of the above-mentioned current situation, an object of the present invention is to provide a technology for providing new information based on presentation materials. [Means for solving the problem]

[0006] [1] An estimation program that causes a computer to function as an acquisition unit, an estimation unit, and an output unit, wherein the acquisition unit acquires material information indicating information contained in presentation materials used in corporate activities, the estimation unit performs input to an estimation model based on the material information, obtains an estimation result regarding the outcome of the corporate activities as an output of the estimation model, and the output unit performs output based on the estimation result.

[0007] This configuration makes it possible to estimate the results of the business activities that used the presentation materials based on the presentation materials, thereby providing information on the results of the business activities, which is the ultimate goal, rather than the quality of the presentation materials themselves, thereby providing more effective support to users.

[0008] [2] The estimation program described in [1] further comprises an evaluation unit, which determines an evaluation for each of a plurality of evaluation items based on the document information, and the output unit further outputs based on the evaluation for each of the evaluation items.

[0009] This configuration provides an evaluation for each specific evaluation item, which can be used to more effectively improve presentation materials.

[0010] [3] The estimation program described in [2], wherein the estimation model is a trained model created by machine learning that takes multiple types of document information as input and outputs estimation results regarding corporate activities, and the evaluation items are items corresponding to the document information that have a large contribution to the estimation results by the estimation model.

[0011] By adopting such a configuration, elements that are thought to have a large impact on the results of business activities can be evaluated as separate evaluation items, and the results can be fed back.

[0012] [4] The estimation program according to [2] or [3], wherein the evaluation unit determines an overall evaluation based on the estimation result by the estimation unit and the evaluation for each evaluation item, and the output unit outputs the overall evaluation.

[0013] This configuration makes it possible to provide a comprehensive evaluation of the entire presentation material, enabling output that is easier to understand.

[0014] [5] The estimation program described in [4], wherein the evaluation unit inputs a generation command including the estimation result by the estimation unit and an evaluation for each evaluation item into a generative model, and determines the overall evaluation based on the output of the generative model.

[0015] With this configuration, for example, a conversational sentence generation model can be used to provide a comprehensive evaluation based on the generation model.

[0016] [6] An estimation program according to any one of [2] to [5], wherein the evaluation items include numerical evaluations based on values ​​identified from the document information.

[0017] [7] An estimation program according to any one of [2] to [6], wherein the evaluation items include a market evaluation regarding the relationship between the business activities and the market.

[0018] [8] The estimation program according to any one of [2] to [7], wherein the evaluation items include an evaluation of the number of explanatory items to be included in the presentation materials.

[0019] [9] An estimation program according to any one of [1] to [8], wherein the estimation result includes the success rate of the business activity.

[0020] By adopting such a configuration, it is possible to provide feedback that is easier to understand and more memorable.

[0021]

[10] An estimation program described in any of [1] to [9], further comprising a memory unit and a judgment unit, wherein the memory unit stores information identifying a trained model corresponding to an activity type indicating the type of the corporate activity, the judgment unit judges the activity type based on the document information, and the estimation unit uses, as the estimation model, a trained model corresponding to the activity type from among a plurality of trained models based on the judgment result by the judgment unit.

[0022] This configuration enables estimations that correspond to various corporate activities. Furthermore, by selecting an appropriate trained model using the judgment model, the user can reduce the effort required to select a trained model themselves.

[0023]

[11] The estimation program described in

[10] , wherein the judgment unit inputs material information about the presentation material to be evaluated into a judgment model created from learning data consisting of pairs of material information and activity types, and judges the activity type based on the output of the judgment model.

[0024]

[12] The estimation program described in

[11] , wherein the judgment model is a model trained as a multi-label classification task that determines multiple labels for one input data, and outputs multiple activity types using document information as input. The judgment unit identifies, from among the activity types output by the judgment model, a type for which a corresponding trained model is stored in the memory unit as the activity type related to the presentation material to be evaluated.

[0025] This configuration allows for a mixture of activity types for which a corresponding trained model exists and activity types for which no trained model has been created at first. This makes it possible to accumulate training data while performing estimation using other models when there is insufficient training data, and then create a model corresponding to that activity type once sufficient training data is available.

[0026]

[13] An estimation system comprising an acquisition unit, an estimation unit, and an output unit, wherein the acquisition unit acquires material information indicating information contained in presentation materials for conducting corporate activities, the estimation unit performs input to an estimation model based on the material information, and obtains an estimation result regarding the outcome of the corporate activities as an output of the estimation model, and the output unit performs output based on the estimation result.

[0027]

[14] An estimation method using an estimation system comprising an acquisition unit, an estimation unit, and an output unit, wherein the acquisition unit acquires material information indicating information contained in presentation materials for conducting corporate activities, the estimation unit performs input to an estimation model based on the material information, obtains an estimation result regarding the outcome of the corporate activities as an output of the estimation model, and the output unit performs output based on the estimation result. [Effects of the Invention]

[0028] According to the present invention, it is possible to provide a technique for providing new information based on presentation materials. [Brief explanation of the drawings]

[0029] [Figure 1] FIG. 1 is a block diagram showing the configuration of a system according to an embodiment of the present invention. [Figure 2] FIG. 4 is a diagram showing an example of information stored in a storage unit of the embodiment. [Figure 3] FIG. 2 is a hardware configuration diagram of an information processing apparatus and a terminal device according to the embodiment. [Figure 4] 1 is a flowchart showing an example of a procedure for creating a trained model that can be used as an estimation model in this embodiment. [Figure 5] 1 is a flowchart showing an outline of the procedure for estimation and evaluation by the system of this embodiment. [Figure 6] 10 is a flowchart showing details of an evaluation procedure performed by the system of the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0030] The present invention relates to a technology for estimating the results of a business activity based on presentation materials used in the business activity. In the following embodiments, the description will be given assuming that the success rate of the business activity is estimated.

[0031] The present invention will now be described in more detail with reference to the accompanying drawings, in which preferred embodiments are shown, but which may be embodied in many different forms and are not limited to the embodiments set forth herein.

[0032] For example, in this embodiment, the configuration, operation, etc. of the estimation system will be described, but similar effects can be achieved by a device having similar functions, a method executed by the device, a computer program that causes a computer device to execute the method, etc. The program may be provided as a non-transitory computer-readable recording medium, or may be provided so as to be downloadable from an external server.

[0033] In the following embodiments, the term "unit" may include, for example, a combination of hardware resources implemented by a broadly defined circuit and software information processing that can be specifically realized by these hardware resources. In this embodiment, "information" is represented by, for example, the physical value of a signal value representing voltage or current, the high or low value of a signal value as a binary bit set consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be performed on a broadly defined circuit.

[0034] A circuit in the broad sense is a circuit realized by appropriately combining a circuit, a processor, a memory, etc. For example, it is a circuit including any of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), etc.

[0035] <1.Definition> First, definitions of key terms used in this embodiment will be explained. In this invention, corporate activities refer to any activity carried out by a company, and include various activities carried out by a company, such as fundraising, proposals and sales to customers in business activities, and comprehensive business activities over a specified period of time.

[0036] The above-mentioned corporate activities are classified into various types depending on their purpose, etc. In this embodiment, for example, types of corporate activities are assumed to be fundraising, customer proposals, and stock price improvement (through business activities). Hereinafter, the above-mentioned types of corporate activities will be referred to as "activity types."

[0037] Furthermore, the "success" of a business activity is defined according to the type of activity. More specifically, a business activity can be judged to be "successful" when the goal or purpose of the business activity according to the type of activity is achieved. For example, if the type of activity is fundraising, the business activity can be said to be successful if funds are raised (such as when permission for investment is granted). For example, if the type of activity is making a proposal to a customer, the business activity can be said to be successful if the customer accepts the proposal.

[0038] In addition, in the present invention, "presentation materials" refers to materials for presentations used in corporate activities. A typical example is a slide deck consisting of multiple pages. Materials used in presentations to explain, propose, etc. to parties according to the activity type are particularly considered as presentation materials in the present invention. More specifically, for example, if the activity type is "fundraising," presentation materials that appeal to investors about business plans and prospects for success are considered, and if the activity type is "client proposal," presentation materials for explaining to clients are exemplified.

[0039] Presentation materials also contain a variety of information, such as text, graphs, and diagrams. The various pieces of information contained in presentation materials are referred to as "material information." Metadata obtained from the presentation material data, such as the number of pages, data size, and file name, may also be used as material information.

[0040] <2. Functional configuration> Fig. 1 is a block diagram showing the configuration of an estimation system according to this embodiment. As shown in Fig. 1, the estimation system includes an estimation device 1, a user terminal 2, and a generation device 3. The estimation device 1, the user terminal 2, and the generation device 3 are configured to be able to communicate with each other via a network NW. In this embodiment, the network NW is an IP (Internet Protocol) network, but there are no limitations on the type of communication protocol, the type of network, etc.

[0041] The estimation device 1 includes an acquisition unit 11, a determination unit 12, an estimation unit 13, an evaluation unit 14, an output unit 15, and a storage unit 16. This is a specific implementation of software-based information processing using hardware. Note that these components do not need to be implemented by a single computer, and may be implemented by multiple computers working together.

[0042] The acquisition unit 11 acquires material information indicating information contained in presentation materials used in business activities. More specifically, the acquisition unit 11 acquires the presentation materials by accepting an upload of the presentation materials from the user terminal 2. The acquisition unit 11 then acquires the material information based on the presentation materials.

[0043] Here, material information refers to any information contained in the presentation materials. For example, if the presentation materials include text, the text information representing the text is included in the material information. Furthermore, if the presentation materials include graphs or diagrams, the images themselves may be used as material information, or the content may be interpreted using any method, such as using an existing machine learning model, and numerical values ​​based on the interpretation may be obtained as material information. Furthermore, numerical values ​​may be obtained from the text information.

[0044] The determination unit 12 determines the activity type based on the document information. Then, from among multiple trained models prepared in advance, a trained model corresponding to the activity type is identified as an estimated model. In this embodiment, from among multiple trained models stored in the storage unit 16, a trained model corresponding to the activity type is identified based on estimated model information indicating the correspondence between the activity type and the trained model, and the trained model is identified as an estimated model.

[0045] The estimation unit 13 uses the learned model identified based on the judgment result by the judgment unit 12 as an estimation model, inputs data to the estimation model based on the document information, and obtains an estimation result regarding the results of corporate activities as an output of the estimation model.

[0046] The evaluation unit 14 determines an evaluation for each of the multiple evaluation items based on the document information of the estimation unit 13. In this embodiment, items with high contributions are selected as evaluation items according to the contributions of the input items in the estimation model. In this embodiment, the evaluation items include numerical evaluation, market evaluation, and excess / deficiency evaluation.

[0047] Numerical evaluation is an evaluation of numerical values ​​specified by document information, such as an evaluation of a company's capital amount. Market valuation is an evaluation of the relationship between a company's activities and the market, and is conducted using publicly available information from other companies. For example, it may include a comparison with market prices, the presence or number of competitors, and whether or not there is overlap with other companies' service (product) concepts. The deficiency / excess evaluation is an evaluation of the amount of explanatory items included in the presentation materials. For example, one possible method is to define the necessary items for each activity type in advance, compare them with the document information, and identify any deficiencies in each item. Also, by separately defining items that are preferable not to be included, it is possible to evaluate excess items as well.

[0048] Furthermore, the evaluation unit 14 of this embodiment determines an overall evaluation based on the estimation result by the estimation unit 13 and the evaluation for each evaluation item. The evaluation unit 14 generates a generation command including the estimation result by the estimation unit 13 and the evaluation for each evaluation item as a sentence in natural language, and transmits it to the generation device 3. The transmission of the generation command to the generation device 3 corresponds to the input to the generative model in the present invention. Then, the generation device 3 transmits the output of the generative model to the estimation device 1, and the evaluation unit 14 acquires it. The evaluation unit 14 then determines an overall evaluation based on the output of the generative model.

[0049] The method for determining the overall rating is not limited to the method using the external generation device 3 described above, but the overall rating may be created by substituting the estimation results by the estimation unit 13 and the ratings for each evaluation item into sentences stored in advance in the storage unit 16. The overall rating may also be created by combining these.

[0050] The output unit 15 performs output based on the estimation result by the estimation unit 13. In this embodiment, the output is performed by displaying the estimation result by the estimation unit 13, the evaluation for each evaluation item determined by the evaluation unit 14, and the overall evaluation, and transmitting the processing results to the user terminal 2. The user terminal 2 displays on a display in accordance with the display processing results.

[0051] The storage unit 16 stores information for identifying a trained model corresponding to an activity type. FIG. 2 is a diagram showing an example of estimation model information stored in the storage unit 16 in this embodiment. In this way, one trained model is identified for each activity type by the estimation model information, and the storage location of the model is stored. That is, there is a one-to-one correspondence between activity types and corresponding trained models. Furthermore, in this embodiment, the storage unit 16 stores a trained model corresponding to each activity type. Note that the trained models may be stored in an external database or the like.

[0052] However, it is not necessary that a trained model exists for all activity types. In reality, there may be activity types for which a trained model cannot be created due to limitations on the amount of training data. In this case, the estimation model information stores information indicating that there is no trained model corresponding to that activity type.

[0053] <3. Hardware configuration> Next, the hardware configuration of the estimation system according to this embodiment will be described. One or more information processing devices 10 (computer devices), such as a general-purpose server or a personal computer, can be used as the estimation device 1 and the generation device 3. Furthermore, a terminal device 9 (computer device), such as a personal computer, a smartphone, or a tablet terminal, can be used as the user terminal 2. In this embodiment, the estimation device 1 is an information processing device 10 in which a computer program (estimation program) that executes the estimation method is installed.

[0054] Fig. 3(a) is a hardware configuration diagram of the information processing device 10. As shown in Fig. 3, the information processing device 10 has a control unit 101, a storage unit 102, and a communication unit 103, which are used to perform the functions of each unit and each process.

[0055] The control unit 101 has a processor such as a CPU that can execute an instruction set, and executes an OS and programs. The storage unit 102 includes a volatile memory such as a RAM capable of storing an instruction set, and a non-volatile recording medium such as an HDD or SSD capable of recording an OS, an estimation program, a DBMS, and the like. The communication unit 103 has an interface for physically connecting to a network, and controls communication with the network NW to input and output information.

[0056] Fig. 3(b) is a hardware configuration diagram of the terminal device 9. As shown in Fig. 2, the terminal device 9 has a control unit 901, a storage unit 902, a communication unit 903, an input unit 904, and an output unit 905, which are used to perform the functions of each unit and each process.

[0057] The control unit 901 has a processor such as a CPU that can execute an instruction set, and executes an OS, application programs, and the like. The storage unit 902 includes a volatile memory such as a RAM capable of storing an instruction set, and a non-volatile recording medium such as an HDD or SSD capable of recording an OS, any application program, and the like. The communication unit 903 has an interface for physically connecting to a network, and controls communication with the network NW to input and output information. The input unit 904 includes an operation input device capable of input processing, such as a touch panel or keyboard, and an audio input device capable of audio input, such as a microphone. The output unit 905 includes a display device capable of display processing, such as a display, and an audio output device, such as a speaker.

[0058] <4. Estimation model creation procedure> Next, we will explain how to create various trained models used as estimation models. Figure 4 is a flowchart showing an example of the procedure for creating a trained model. The procedure shown in Figure 4 is for creating one trained model, but different trained models are created for each activity type using this procedure.

[0059] First, in step S11, learning data is acquired. The learning data is a set of multiple types of document information in presentation materials and the success or failure of activities carried out using the presentation materials. In step S11, a data set including multiple such learning data is acquired.

[0060] For example, if the activity type is "fundraising," text information can be obtained from descriptions such as "mission" and "vision" included in presentation materials and used as input for the model. In this case, information on whether fundraising was successful or not is used to indicate the success or failure of the activity. Such document information obtained from presentation materials used in past corporate activities and the results of those corporate activities are obtained as learning data.

[0061] Next, in step S12, a learning process is executed. Specifically, learning data corresponding to various activity types is provided to the machine learning model, and the model parameters are adjusted so that the model output matches the learning data, thereby performing learning. Any known machine learning method can be used, and those skilled in the art will be able to understand the specific procedure, so a description thereof will be omitted.

[0062] When the learning process is completed, in step S13, the learned model is analyzed to analyze the contribution of each input item (document information). In machine learning, the calculations between input and output are usually a black box, but there are known methods for analyzing the extent to which multiple input items each contributed to determining the output.

[0063] For example, PDP, LIME, SHAP, etc. are known as methods for calculating the degree to which each feature contributes to a predicted value. These methods are widely used in the field of machine learning and can be understood by those skilled in the art, so detailed explanations will be omitted. In step S13, the contribution of each input item is analyzed using any of these known methods.

[0064] Then, in step S14, the input items are ranked in descending order of the degree of contribution analyzed in step S13. By ranking the items according to their degree of contribution in this way, it is possible to confirm which items in the presentation materials are likely to affect the success or failure of business activities.

[0065] In this embodiment, the items (document information) with higher rankings are evaluated separately as evaluation items according to the ranking determined in step S14. This makes it possible to individually evaluate and provide feedback on document information that is likely to affect the success or failure of business activities, thereby providing more useful information.

[0066] <5. Overall flow> The overall procedure for performing estimation and evaluation based on presentation materials using the trained model created as described above will be described with reference to Fig. 5. First, in step S21, the acquisition unit 11 acquires the presentation materials from the user terminal 2.

[0067] Then, in step S22, the acquisition unit 11 analyzes the presentation materials and acquires the material information. Specifically, if text information is included, it may be acquired as the material information, or if the presentation materials are provided in the form of images, the text information may be acquired using character recognition technology such as OCR (Optical Character Recognition / Reader).

[0068] In step S23, the determination unit 12 determines the activity type corresponding to the presentation material based on the material information acquired by the acquisition unit 11. More specifically, the determination unit 12 inputs the material information acquired by the acquisition unit 11 into a determination model created from learning data consisting of pairs of material information and corresponding activity types, and obtains the activity type as an output of the determination model.

[0069] The items of material information input here may be different from the information input to the estimation model described below. For example, in the judgment model, keywords and summaries can be generated based on the text information of the presentation materials and used as input to the judgment model. On the other hand, the input information to the estimation model can be, for example, the planned procurement amount, operation status, service price, etc., and may differ depending on the corresponding activity type.

[0070] Similarly to the estimation model, the determination model is created by acquiring multiple pieces of training data and executing a training process. Note that the determination model does not require contribution analysis. The procedure for the training process may be determined arbitrarily, and those skilled in the art will be able to understand the details, so a detailed explanation will be omitted.

[0071] The classification model learns multi-label classification, which determines multiple labels for a single input data. That is, the classification model outputs multiple labels, i.e., activity types, for the input. Specifically, the classification model outputs a confidence score for each activity type, and presents activity types with confidence scores equal to or greater than a predetermined threshold in descending order of confidence score.

[0072] Then, the estimation model information stored in the storage unit 16 is referenced to confirm the trained model corresponding to the activity type. If a trained model corresponding to the activity type exists (is registered), the trained model is identified as the estimation model to be used for estimation, and the activity type is identified as the activity type related to the presentation material to be evaluated. Note that if there are multiple activity types for which trained models exist, the trained model corresponding to the activity type with the highest reliability score is identified as the estimation model to be used for estimation.

[0073] In step S24, an evaluation process is performed by the estimation unit 13 and the evaluation unit 14. Details of the evaluation process will be described later.

[0074] As a result of the evaluation process, an estimation result by the estimation unit 13, an evaluation result for each evaluation item by the evaluation unit 14, and an overall evaluation by the evaluation unit 14 are obtained. In step S25, the output unit 15 performs display processing on these pieces of information and transmits the processing result to the user terminal 2, thereby causing the user terminal 2 to display the content thereof.

[0075] <6. Evaluation process details> Details of the evaluation process in step S24 in Fig. 5 will be described below with reference to Fig. 6. Note that, when the estimation model uses information other than the document information used by the determination model, the acquisition unit 11 further acquires document information used by the estimation model before the procedure shown in Fig. 6.

[0076] 6 vary depending on the activity type corresponding to the presentation material. Therefore, the storage unit 16 stores, for each activity type, evaluation items for evaluation other than estimation using the estimation model. In the above explanation, the determination unit 12 determines the activity type in order to determine the estimation model, but the activity type is also referenced when performing evaluation for each evaluation item described below.

[0077] First, in step S31, the evaluation unit 14 determines a numerical evaluation as an evaluation item. The numerical evaluation is an evaluation based on the numerical values ​​contained in the document information. The target of the numerical evaluation is an item corresponding to the document information that has a large contribution to the estimation result by the estimation model. For example, in an estimation model corresponding to the activity type of "fund raising," if the document information on "capital amount" has a large contribution to the estimation result, a numerical evaluation is performed on the capital amount.

[0078] In step S32, the evaluation unit 14 determines a market valuation as an evaluation item. The market valuation is an evaluation of the relationship between a company's activities and the market. For example, when performing an evaluation corresponding to the activity type of "fundraising," it is assumed that the evaluation unit 14 refers to public information of other companies in the market, compares the price included in the document information with the average market price, and determines whether it is higher or lower.

[0079] The numerical evaluation and market evaluation may be performed using a model created by machine learning, or may be performed by simple rule-based numerical comparison. When using a model created by machine learning, any evaluation may be performed, such as numerical estimation or strength determination using a language model.

[0080] In step S33, the evaluation unit 14 determines an excess / deficiency evaluation as an evaluation item. The market evaluation is an evaluation of the excess / deficiency of explanatory items that should be included in the presentation materials. For example, the memory unit 16 stores necessary explanatory items and explanatory items that are preferable not to be included for each activity type, and the evaluation unit 14 refers to this and evaluates whether there is an excess or deficiency. If there is an excess or deficiency, it is expected that the specific excess and deficiency items will be determined as the excess / deficiency evaluation.

[0081] Then, in step S34, the estimation unit 13 estimates the results of the business activity by inputting the document information into the estimation model determined by the determination unit 12. In this embodiment, the estimation model outputs the success or failure of the business activity and its probability. The probability is the success rate of the business activity.

[0082] In this way, it is possible to estimate the numerical evaluation, market evaluation, excess / deficiency evaluation, and success rate of business activities. Then, in step S35, the evaluation unit 14 inputs this information into the generative model and obtains an overall evaluation as the output of the generative model. In this embodiment, input to the generative model is performed by sending a generation command including the numerical evaluation, market evaluation, excess / deficiency evaluation, and success rate of business activities to the generation device 3. Then, an overall evaluation comment is obtained in text form as the output of the generative model.

[0083] The above-described procedure is merely an example, and the order may be changed as desired within the scope of possibility. For example, steps S31 to S34 may be performed in any order, and may be executed in parallel.

[0084] As described above, according to this embodiment, an estimation model can be determined based on presentation materials, and then an estimation can be made regarding the success or failure of activities. As a result, a user can obtain an estimation regarding the results of corporate activities using the presentation materials and an evaluation for each evaluation item corresponding to the elements that contribute to the results of the corporate activities, simply by uploading the presentation materials via the user terminal 2.

[0085] Furthermore, it is possible to show the user specific evaluations from different perspectives, such as numerical evaluation, market evaluation, and excess / deficiency evaluation, making it possible to provide more specific information for improving presentation materials. [Explanation of symbols]

[0086] 1: Estimation device 11: Acquisition part 12: Judgment section 13: Estimation part 14: Evaluation section 15: Output section 16: Storage section 2: User terminal 3:Generation device 10: Information processing device 101: Control unit 102: Storage section 103: Communications Department 9: Terminal device 901: Control unit 902: Storage section 903: Communications Department 904: Input section 905: Output section

Claims

1. An estimation program that causes a computer to function as an acquisition unit, an estimation unit, and an output unit, the acquiring unit acquires material information indicating information included in presentation materials used in business activities; the estimation unit performs input to an estimation model based on the document information, and obtains an estimation result regarding the outcome of the corporate activity as an output of the estimation model; The output unit performs an output based on the estimation result.

2. The apparatus further includes an evaluation unit, determining an evaluation for each of a plurality of evaluation items based on the document information; The estimation program according to claim 1 , wherein the output unit further performs output based on the evaluation for each of the evaluation items.

3. The estimation model is a trained model created by machine learning that takes multiple types of document information as input and outputs estimation results regarding corporate activities, The estimation program according to claim 2 , wherein the evaluation items are items corresponding to the document information that have a large contribution to the estimation result by the estimation model.

4. The evaluation unit determining a comprehensive evaluation based on the estimation result by the estimation unit and the evaluation for each of the evaluation items; The estimation program according to claim 2 , wherein the output unit outputs the overall evaluation.

5. The estimation program according to claim 4 , wherein the evaluation unit inputs a generation command including the estimation result by the estimation unit and an evaluation for each evaluation item into a generative model, and determines the overall evaluation based on the output of the generative model.

6. The estimation program according to claim 2 , wherein the evaluation items include numerical evaluations based on values ​​identified from the document information.

7. The estimation program according to claim 2 , wherein the evaluation items include a market evaluation regarding a relationship between the business activities and the market.

8. The estimation program according to claim 2 , wherein the evaluation items include an excess / deficiency evaluation regarding an excess or deficiency of explanatory items to be included in the presentation materials.

9. The estimation program according to claim 1 , wherein the estimation result includes a success rate of the business activity.

10. The device further includes a storage unit and a determination unit, The storage unit stores information for identifying a trained model corresponding to an activity type indicating a type of the corporate activity, The determination unit determines the activity type based on the material information, The estimation program according to claim 1 , wherein the estimation unit uses, as the estimation model, a trained model corresponding to the activity type from among a plurality of trained models based on a determination result by the determination unit.

11. The estimation program according to claim 10, wherein the determination unit inputs material information about the presentation material to be evaluated into a determination model created from learning data consisting of pairs of material information and activity types, and determines the activity type based on the output of the determination model.

12. The determination model is a model trained as a task of multi-label classification, which determines multiple labels for one input data, and receives document information as input and outputs multiple activity types; The estimation program according to claim 11, wherein the judgment unit identifies, among the activity types output by the judgment model, a type for which a corresponding trained model is stored in the memory unit as the activity type related to the presentation material to be evaluated.

13. An estimation system including an acquisition unit, an estimation unit, and an output unit, the acquiring unit acquires material information indicating information included in presentation materials for conducting business activities; the estimation unit performs input to an estimation model based on the document information, and obtains an estimation result regarding the outcome of the corporate activity as an output of the estimation model; The output unit performs output based on the estimation result.

14. An estimation method using an estimation system including an acquisition unit, an estimation unit, and an output unit, the acquiring unit acquires material information indicating information included in presentation materials for conducting business activities; the estimation unit inputs data into an estimation model based on the document information, and obtains an estimation result regarding the outcome of the corporate activity as an output of the estimation model; The estimation method, wherein the output unit performs output based on the estimation result.

Images