Extraction system, extraction method, and extraction program

Abstract

To provide an extraction system configured to facilitate extraction of a product similar to a target product.SOLUTION: An extraction system 10 includes an output unit 13, an acquisition unit 14, and an update unit 15. The output unit 13 displays a feature amount used for similarity determination by an extraction model that extracts a product similar to a target product, a set value of an importance level of the feature amount, and a change field for changing the set value of the importance level of the feature amount. The acquisition unit 14 acquires a change value for changing the set value of the importance level of the feature amount input in the change field of the output screen. The update unit 15 updates the set value of the importance level of the feature amount in the extraction model based on the change value. The extraction system 10 can be used, for example, to support decision making.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present disclosure relates to extraction systems and the like. [Background technology]

[0002] In the planning stage of a new product, for example, a demand forecast for the new product is made. The demand forecast for the new product is made, for example, using sales performance data of products similar to the new product among products on the market. In addition, the products similar to the new product are selected, for example, by a person who makes the demand forecast for the new product.

[0003] The sales volume forecasting computer in Patent Document 1 extracts products that have similar features to the new product. The sales volume forecasting computer in Patent Document 1 then predicts the sales volume of the new product based on the sales volumes of the similar products. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2022-120816 A Summary of the Invention [Problem to be solved by the invention]

[0005] In the sales volume forecast of Patent Document 1, it may be difficult to extract similar products suitable for the forecast.

[0006] In order to solve the above-mentioned problems, the present disclosure aims to provide an extraction system etc. that can easily extract products similar to a target product. [Means for solving the problem]

[0007] In order to solve the above problems, the extraction system disclosed herein comprises an output means for outputting a screen displaying features used by an extraction model that extracts products similar to a target product for similarity judgment, a feature importance setting value, and a change field for changing the feature importance setting value, an acquisition means for acquiring a change value for changing the feature importance setting value, which is entered into the change field of the output screen, and an update means for updating the feature importance setting value in the extraction model based on the change value.

[0008] The extraction method disclosed herein outputs a screen displaying features used by an extraction model that extracts products similar to a target product for similarity judgment, a setting value for the feature importance, and a modification field for changing the setting value of the feature importance, obtains a modification value for changing the setting value of the feature importance that is entered into the modification field of the output screen, and updates the setting value of the feature importance in the extraction model based on the modification value.

[0009] The extraction program disclosed herein causes a computer to execute the following processes: outputting a screen displaying features used by an extraction model that extracts products similar to a target product for similarity judgment, a setting value of the feature importance, and a change field for changing the setting value of the feature importance; acquiring a change value for changing the setting value of the feature importance that is entered into the change field of the output screen; and updating the setting value of the feature importance in the extraction model based on the change value. Effect of the Invention

[0010] According to the present disclosure, it is possible to easily extract products similar to a target product. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram illustrating an example of a configuration of an information processing system according to the present disclosure. [Diagram 2] FIG. 2 is a diagram showing an example of a display screen in the present disclosure. [Diagram 3] FIG. 2 is a diagram showing an example of a display screen in the present disclosure. [Figure 4]FIG. 1 is a diagram illustrating an example of a configuration of an extraction system according to the present disclosure. [Diagram 5] FIG. 2 is a diagram showing an example of a display screen in the present disclosure. [Figure 6] FIG. 2 is a diagram showing an example of a display screen in the present disclosure. [Figure 7] FIG. 2 is a diagram showing an example of a display screen in the present disclosure. [Figure 8] FIG. 2 is a diagram illustrating an example of an operation flow of the extraction system according to the present disclosure. [Figure 9] FIG. 2 is a diagram illustrating an example of an operation flow of the extraction system according to the present disclosure. [Figure 10] FIG. 2 is a diagram illustrating an example of an operation flow of the extraction system according to the present disclosure. [Figure 11] FIG. 2 is a diagram illustrating an example of a hardware configuration according to the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] An embodiment of the present disclosure will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an example of a configuration of an information processing system. The information processing system includes, for example, an extraction system 10, a terminal device 20, and a data management device 30. The extraction system 10 is connected to the terminal device 20, for example, via a network. The extraction system 10 is connected to the data management device 30, for example, via a network. There may be a plurality of terminal devices 20 and a plurality of data management devices 30.

[0013] The information processing system is, for example, a system that extracts products similar to a target product using an extraction model. The extraction model is, for example, a learning model that uses attribute data of the target product as input and extracts products similar to the target product. Products may include goods and services. Products may also include agricultural and marine products. Products are not limited to the above.

[0014] The target product is, for example, a product for which a user wishes to obtain information on a product similar to the target product. A product similar to the target product is, for example, a product to be compared with the target product. A product similar to the target product is, for example, a product to be used as a reference when considering commercialization of the target product. For example, when predicting the sales volume of a new product, a demand forecast for the new product is made using data on the sales volume of a product similar to the new product. In this case, the target product is a new product. A product similar to the new product is, for example, a product with similar characteristics to the new product. The sales volume of the new product is likely to show a similar trend to the actual sales volume of a product with similar characteristics to the new product. For this reason, information on a product with similar characteristics to the new product can be used, for example, to predict the sales volume of the new product. In addition, a product similar to the target product may be a product that competes with the target product due to its similar characteristics when the target product is sold in the market.

[0015] Furthermore, a product similar to the target product is, for example, a product whose attribute data is similar to that of the target product. An attribute is, for example, data that represents the characteristics of a product. An attribute is data that represents the characteristics of a product using separation or numerical values. The attribute data is, for example, data related to attributes of a product that may affect the use of the extraction result by the extraction model, among the attributes of the product. For example, when the use of the extraction result is a prediction of sales volume, the attribute data is data that may affect the prediction of sales volume. The attribute data is, for example, one or more items of design, price, target customer base, sales season, sales volume, sales region, distribution channel, sales form, designer, manufacturer, seller, brand, and place of origin. The attribute data is not limited to the above. Furthermore, the target product is not limited to a new product. For example, the target product is a product in the planning stage, a product in the concept stage, or a product for market research. The target product may be a product that has already been released on the market.

[0016] The extraction of products similar to the target product can be performed for products related to, for example, clothing, shoes, bags, watches, ornaments, electrical appliances, electronic devices, automobiles, bicycles, sporting goods, beverages, food, furniture, fixtures, or housing. The types of products to be extracted are not limited to the above.

[0017] FIG. 2 is an example of a display screen that displays the results of extracting products similar to the target product. In the example of the display screen in FIG. 2, three products similar to the target product are displayed. In addition, in the example of the display screen in FIG. 2, attribute data indicating the characteristics of each product is displayed. For example, by referring to the screen shown in the example of the display screen in FIG. 2, the user can check the products similar to the target product and check the validity of the extraction results. Details of FIG. 2 will be described later.

[0018] Furthermore, in the case where the extraction result of a product similar to the target product is used for product demand forecasting, the product demand forecasting is, for example, a forecast of the demand amount of the product. The demand amount is, for example, the sales quantity or sales amount of the product. The demand amount is not limited to the above. The demand forecasting is performed, for example, using a demand forecasting model based on the attribute data of the product. The demand forecasting model is generated, for example, by learning the relationship between the attribute data of the product and the demand of the product. The data of the product similar to the target product is used, for example, as learning data for the demand forecasting model. Furthermore, the data of the product similar to the target product is used, for example, to verify the results of the demand forecasting by the demand forecasting model. For example, the validity of the results of the demand forecasting can be verified by comparing the attribute data of the product similar to the target product and the demand amount of the product similar to the target product with the attribute data of the target product and the demand amount of the target product predicted by the demand forecasting model.

[0019] FIG. 3 shows an example of a display screen in a product demand forecasting system in which an extraction result of a product similar to a target product is used. The example of the display screen in FIG. 3 is an example of a screen that displays a forecast result of the sales volume of product X as a result of product demand forecast. In the example of the display screen in FIG. 3, a graph showing the forecast value of the sales volume of product X and the sales performance of product A-1 is displayed. The sales performance of product A-1 is the past sales volume performance data for each month. In addition, in the example of the display screen in FIG. 3, the extraction result of a product similar to the target product is displayed. In addition, in the example of the display screen in FIG. 3, for example, information of items similar to the example of the display screen in FIG. 2 is displayed as an extraction result of a product similar to the target product. In the example of the display screen in FIG. 3, by clicking an image of a product similar to the target product, data of the sales performance of the product similar to the target product is displayed in a graph.

[0020] Here, a specific example of the configuration of the extraction system 10 will be described. Fig. 4 is a diagram showing an example of the configuration of the extraction system 10. The extraction system 10 basically includes an output unit 13, an acquisition unit 14, and an update unit 15. The extraction system 10 also includes, for example, an extraction unit 11, a model update unit 16, a name generation unit 17, and a storage unit 12.

[0021] The extraction unit 11 extracts products similar to the target product using, for example, an extraction model. The extraction model is, for example, a learning model that extracts products similar to the target product. A method for generating an extraction model will be described later. The extraction model, for example, uses information about the target product as an input of the extraction model and outputs products similar to the target product as an extraction result. The extraction unit 11 extracts products similar to the target product from, for example, data stored in the data management device 30. When multiple extraction models are generated, the extraction unit 11 extracts products similar to the target product using an extraction model selected on a screen output by the output unit 13. In addition, when a setting value of the importance of a feature is changed, the extraction unit 11 extracts products similar to the target product using an extraction model based on the changed setting value. In addition, when the extraction model is a learning model capable of estimating the reason for the extraction result, the extraction unit 11 may generate the reason for the extraction result based on the output of the extraction model.

[0022] The extraction unit 11, for example, extracts products whose similarity is equal to or greater than a standard as products similar to the target product. The standard for the similarity to be extracted as a similar product is set based on a value at which a product can be considered to be similar to the target product when the product's similarity is equal to or greater than the standard value. The similarity is, for example, an index indicating the degree to which the target product and the extracted product are similar. The similarity is, for example, the accuracy of classification by the extraction model. The extraction unit 11 may extract products whose similarity is from the top to a predetermined rank as products similar to the target product. The predetermined rank is set so that the number of extracted products is appropriate for a user to refer to.

[0023] The extraction of a product similar to the target product may be performed by using any one of a plurality of extraction models that have been generated. When a plurality of extraction models have been generated, the extraction unit 11 extracts a product similar to the target product using, for example, an extraction model selected by a user. The extraction unit 11 may also extract a product similar to the target product using an extraction model based on the type of the target product. The extraction unit 11 may also extract a product similar to the target product using an extraction model based on the user among the plurality of extraction models. For example, when the user is a person in the food industry, the extraction unit 11 may extract a product similar to the target product using an extraction model generated for the food industry. The extraction unit 11 may extract a product similar to the target product using an extraction model according to the affiliation of the user based on, for example, the login information of the user. In this case, the relationship between the affiliation of the user and the extraction model is set, for example, as a table showing the relationship between the affiliation of the user and the extraction model.

[0024] The memory unit 12 stores, for example, data related to a process for extracting products similar to the target product. The memory unit 12 stores, for example, an extraction model. The memory unit 12 stores, for example, a setting value of the importance of a feature in the extraction model. The memory unit 12 stores, for example, the weight of the feature in the extraction model. When the setting value of the importance of a feature is changed, the memory unit 12 may store a change history of the setting value. The memory unit 12 stores, for example, a name generation model. The memory unit 12 stores, for example, the extraction result of a product similar to the target product by the extraction model. The memory unit 12 also stores learning data used for learning and re-learning the extraction model.

[0025] The output unit 13 outputs a screen displaying the feature amount used by the extraction model for similarity judgment, the set value of the feature amount importance, and a change field for changing the set value of the feature amount importance. The extraction model is a learning model for extracting products similar to the target product. The output unit 13 outputs, for example, the feature amount in the extraction model stored in the storage unit 12 and the set value of the feature amount importance. The output unit 13 outputs, for example, to the terminal device 20, a screen displaying the feature amount used by the extraction model for extracting products similar to the target product for similarity judgment, the set value of the feature amount importance, and a change field for changing the set value of the feature amount importance. The output unit 13 may output, for example, to a display device (not shown) connected to the extraction system 10, a screen displaying the feature amount used by the extraction model for extracting products similar to the target product for similarity judgment, the set value of the feature amount importance, and a change field for changing the set value of the feature amount importance.

[0026] The feature amount is, for example, a variable related to the feature of the product, and is a variable used by the extraction model for the similarity judgment. The feature amount weight is, for example, a weight for each feature amount used by the extraction model for the similarity judgment. The feature amount importance is an index obtained by converting the feature amount weight in the extraction model to an index that is easy for the user to understand. The feature amount importance is, for example, a value obtained by standardizing the weight of each feature amount used by the extraction model for the similarity judgment. By standardizing the feature amount weight, for example, the user can easily change the setting value related to the feature amount weight. The feature amount importance may be a coefficient multiplied by the weight of each feature amount used by the extraction model for the similarity judgment. The feature amount importance is not limited to the above. When the feature amount importance is a value obtained by standardizing the feature amount weight, the feature amount importance is calculated by standardizing the weight of each feature amount using the weight value of a predetermined feature amount. The predetermined feature amount is, for example, a feature amount that is expected to be important in the similarity judgment. The predetermined feature amount may be a feature amount with the largest weight value. The predetermined feature amount is not limited to the above.

[0027] When a change value for changing the setting value of the importance of a feature is input into the change field of the output screen by a user operation, the output unit 13 outputs, for example, a screen displaying the setting value of the importance of the feature after the change. The change value for changing the setting value of the importance of a feature is acquired by the acquisition unit 14, for example, from the terminal device 20. When a change value for changing the setting value of the importance of a feature is input into the change field of the output screen by a user operation, the output unit 13 may output a screen displaying the setting values ​​of the importance of the feature before and after the change.

[0028] The output unit 13 may output a screen displaying the feature importance setting value using a predetermined numerical value indicating a standard value when the feature importance setting value has not been changed from the setting value in the extraction model. For example, when the predetermined numerical value indicating the standard value is set to 3, the output unit 13 outputs the feature importance setting value that has not been changed from the setting value in the extraction model as 3. The predetermined numerical value is set so that the feature importance setting value can be increased or decreased. When a change value for changing the feature importance setting value is input by a user's operation into the change field of the output screen, the output unit 13 outputs a screen displaying the feature importance setting value using the input change value. The change value for changing the feature importance setting value is acquired by the acquisition unit 14, for example, from the terminal device 20.

[0029] For example, when outputting a screen displaying a product extracted by the extraction model, the output unit 13 outputs a screen further displaying a feature amount that contributed to the similarity judgment with the target product. The feature amount that contributed to the similarity judgment with the target product may be displayed on the same screen as the screen displaying the product extracted by the extraction model. By referring to the feature amount that contributed to the similarity judgment with the target product, the user can recognize the degree of influence of the feature amount on the similarity judgment. For example, when extracting a product similar to the target product for use in product demand forecasting, the user can determine whether the extracted product is appropriate as a product similar to the target product by referring to the feature amount that contributed to the similarity judgment with the target product. In addition, the user can more appropriately set the importance of the feature amount by, for example, changing the setting value of the importance of the feature amount by referring to the feature amount that contributed to the similarity judgment with the target product.

[0030] The output unit 13 outputs, for example, a screen displaying the feature amount that contributed most to the similarity judgment with the target product. The output unit 13 outputs, for example, a screen displaying the feature amounts that contributed to the similarity judgment by the extraction model in the order of the contribution to the similarity judgment. The output unit 13 outputs, for example, a screen displaying feature amounts that have a contribution degree to the similarity judgment with the target product from the top to a predetermined rank. The output unit 13 may also output a screen displaying feature amounts whose contribution degree to the similarity judgment with the target product is equal to or greater than a predetermined criterion. The predetermined rank and the predetermined criterion are set, for example, so that the number of feature amounts to be output is an appropriate number for the user to refer to the displayed feature amounts. In addition, when the extraction model is a learning model capable of estimating the reason for the extraction result, the output unit 13 may output a screen displaying the reason for the extraction together with the product extracted by the extraction model.

[0031] When multiple products similar to the target product are displayed, the output unit 13 may output a screen that allows a product displayed on the screen among the multiple products to be changed by scrolling the screen. Also, when multiple products similar to the target product are displayed, the output unit 13 may output a screen that displays a button that allows a product to be displayed to be changed by switching pages. The form of displaying multiple products similar to multiple target products may be set appropriately.

[0032] The output unit 13 may output, for example, a screen for selecting a product to be used as learning data from candidates for products to be used as learning data for the extraction model. The output unit 13 may output, for example, a screen for displaying products extracted by the extraction model as candidates for learning data. The output unit 13 may also output a screen for further displaying feature amounts used by the extraction model to extract products as candidates for learning data. The output unit 13 may output, for example, feature amounts with weight values ​​equal to or greater than a reference value among the feature amounts used by the extraction model to extract products. The reference value of the weight value is set, for example, based on a value that is considered to contribute to the similarity judgment made by the extraction model. The output unit 13 may further output, for example, a screen for displaying products extracted by the extraction model and feature amounts used by the extraction model to extract products as candidates for learning data.

[0033] The output unit 13 outputs, for example, a screen for selecting a product to be used as learning data from among the products extracted by the extraction model. That is, the output unit 13 outputs, for example, a screen for selecting a product to be used for re-learning the extraction model. The output unit 13 outputs, for example, a screen displaying the product extracted by the extraction model and a checkbox to be checked when selecting to use the product as learning data. The output unit 13 also outputs, for example, a screen on which it is possible to select whether or not to use the feature amount used by the extraction model to extract the product as learning data. The output unit 13 outputs, for example, a screen displaying the feature amount used by the extraction model to extract the product and a checkbox to be checked when selecting to use the feature amount as learning data.

[0034] When an extraction model selected by a user from among multiple extraction models is used to extract a product similar to the target product, the output unit 13 outputs, for example, a screen for selecting one of the multiple extraction models.The output unit 13 then outputs a screen displaying the feature amounts and feature importance setting values ​​in the selected extraction model, and a change field for changing the feature importance setting value.The output unit 13 outputs, for example, a screen displaying a pull-down menu for selecting an extraction model from the multiple extraction models.

[0035] The output unit 13 outputs, for example, buttons that increase or decrease the set value when operated on the screen as a change field for changing the set value of the importance of the feature. The output unit 13 outputs, for example, an image for increasing or decreasing the set value when a "+" button and a "-" button are operated on the screen. The buttons may include a slide bar. The output unit 13 outputs, for example, a slide bar that increases or decreases the set value when operated. The change field for changing the set value of the importance of the feature is not limited to the above.

[0036] The output unit 13 may output a screen that displays buttons for transitioning between a display screen of the extraction results and a screen for changing the setting value. The output unit 13 may also output a screen that displays buttons for transitioning between a display screen of the extraction results and a screen for selecting learning data for the extraction model. The output unit 13 outputs a screen including a button for switching between a screen for displaying products extracted by the extraction model and a screen for displaying a change field for changing the setting value of the importance of the feature in the extraction model, for example. When a button is pressed on the screen, the output unit 13 outputs a screen associated with the pressed button.

[0037] The output unit 13 outputs, for example, a screen further displaying standard feature quantities in the industry in which the target product is distributed. The output unit 13 outputs, for example, a screen displaying feature quantities that are frequently used to extract similar products in the industry in which the target product is distributed. For example, in the case where the color and price of a package have a large impact on sales volume in the industry in which the target product is distributed, the output unit 13 outputs a screen indicating that color and price are important items. The standard feature quantities in the industry in which the target product is distributed are set, for example, as a table indicating the relationship between product types and feature quantities.

[0038] Each of the above-described screens may be generated in the terminal device 20. When image data for a screen is generated in the terminal device 20, the output unit 13 outputs to the terminal device 20 an instruction regarding the display content.

[0039] The acquisition unit 14 acquires a change value of the importance of a feature amount input on the screen. The acquisition unit 14 acquires a change value of the importance of a feature amount input by a user's operation from the terminal device 20, for example.

[0040] The acquisition unit 14 acquires a selection result of a product to be used as learning data, which is input by, for example, an operation of selecting a product from a plurality of products displayed on the screen. The acquisition unit 14 also acquires, for example, a selection of feature amounts to be used as learning data, which is selected on the screen. The acquisition unit 14 may acquire a product to be used as learning data and a selection of feature amounts, which are selected on the screen.

[0041] The acquisition unit 14 acquires, for example, information related to the target product. The information related to the target product is, for example, information that identifies a target product for which a user desires to extract a similar product. The information that identifies the target product is, for example, the name of the target product and attribute data of the target product. The information that identifies the target product is not limited to the above.

[0042] When multiple extraction models have been generated, the acquisition unit 14 acquires, for example, a selection of an extraction model to be used for extracting products similar to the target product. The acquisition unit 14 acquires, for example, a selection of an extraction model to be used for extracting products similar to the target product from the terminal device 20. The selection of an extraction model to be used for extracting products similar to the target product is input, for example, to the terminal device 20 by a user's operation.

[0043] The update unit 15 updates the setting value of the feature importance in the extraction model based on the change value acquired by the acquisition unit 14. Furthermore, the update unit 15 updates the weight of the feature, which is an actual parameter in the extraction model, based on, for example, the changed setting value of the feature importance. For example, when the setting value of the feature importance is standardized, the update unit 15 calculates the weight of the feature from the standardized post-update setting value and updates the weight of the feature in the extraction model.

[0044] The update unit 15 may update the setting value of the importance of the feature in the extraction model as a temporary value based on the change value acquired by the acquisition unit 14. The update unit 15 may update the setting value of the importance based on a selection of whether to update the setting value of the importance as a temporary value or a non-temporary value. Updating as a temporary value means that the updated value is not used when similar products are extracted again after similar products have been extracted once based on the updated setting value. Updating as a non-temporary value means that the updated value is used when similar products are extracted again after similar products have been extracted based on the updated setting value. The update unit 15 may add an identifier to each of the change values ​​acquired by the acquisition unit 14 to update the setting value of the importance of the feature in the extraction model so that there are multiple setting values. The update unit 15 may update the setting value of the importance of the feature in the extraction model based on a selection of which change value to use for the update among multiple change values, for example.

[0045] The model update unit 16 generates an extraction model by, for example, performing re-learning using the selected product or feature amount as learning data. Generating an extraction model by re-learning is also called updating the extraction model. The model update unit 16 performs re-learning using, for example, the same algorithm as that used when generating the extraction model.

[0046] The model update unit 16 generates the extraction model by, for example, inverse reinforcement learning. The model update unit 16 generates the extraction model by, for example, performing inverse reinforcement learning with the attribute data of the product as explanatory variables and the extraction results of similar products as objective variables. The model update unit 16 may also generate the extraction model by inverse reinforcement learning based on factorized asymptotic Bayesian inference. When performing inverse reinforcement learning based on factorized asymptotic Bayesian inference, the model update unit 16 generates a plurality of linear models that are case-separated by rules in the form of a decision tree with the attribute data of the product as explanatory variables and the extraction results of similar products as objective variables. The model update unit 16 then generates the extraction model by sequentially performing the processes of optimizing the case-separation conditions of the data, generating a linear model by optimizing the combination of explanatory variables, and deleting unnecessary linear models. The extraction model generated by such a method of generating a learning model using a combination of different explanatory variables makes it possible to explain the extraction results using case-separation conditions that have a strong influence on the extraction results of similar products, thereby improving the explainability of the extraction results. In other words, the extraction model is a learning model that can explain the reason why a product was extracted as a product similar to the target product.

[0047] The model update unit 16 may generate the extraction model by deep learning using a neural network. The algorithm used for the machine learning to generate the extraction model is not limited to the above.

[0048] The model updating unit 16 stores the updated extraction model in, for example, the storage unit 12. The model updating unit 16 may add an identifier to the updated extraction model and store the updated extraction model separately from the extraction model before the update. For example, the model updating unit 16 adds a name to the updated extraction model that is different from the name before the update and stores the updated extraction model separately from the extraction model before the update.

[0049] When generating an extraction model in the extraction system 10, the model update unit 16 generates, for example, an extraction model that extracts products similar to a target product. The model update unit 16 generates an extraction model by learning, for example, the relationship between product attribute data and products selected as similar products. The extraction model is a learning model that extracts products similar to a target product from data of the target product. The extraction model may also be generated in a system external to the extraction system 10.

[0050] When generating an extraction model in the extraction system 10, a data acquisition means (not shown) may acquire learning data of the extraction model. The data acquisition means acquires, for example, a product selected as a product similar to the target product and attribute data of the product as learning data of the extraction model. The data acquisition means acquires, for example, a product selected by an expert in the past as a product similar to the target product and attribute data of the product as learning data. The data acquisition means acquires, for example, a product selected by a demand forecasting expert as a product similar to the target product of the demand forecast and attribute data of the product as learning data. The learning data acquired by the data acquisition means may be a product selected by a data scientist or a user of the extraction system 10 and attribute data of the product.

[0051] For example, when an extraction model is updated, the name generation unit 17 generates a name for the extraction model based on the features of the learning data. Also, for example, when a setting value of the importance of a feature is updated, the name generation unit 17 generates a name for the extraction model based on the changed value of the setting value.

[0052] The name generating unit 17 generates a name of the extraction model based on, for example, the attributes of the learning data and the feature quantities with high importance. The relationship between the attributes of the learning data and the feature quantities with high importance and the name of the extraction model is set, for example, as a table. For example, when the season in which the product is sold is important in the similarity judgment of the extraction model, the attribute "season" and the name "seasonal emphasis" are associated in the table. In such a case, the name generating unit 17 generates a name of the extraction model such as "seasonal emphasis model". In addition, the relationship between the change value of the setting value and the name of the extraction model may be set as a table. For example, when the importance of the season in which the product is sold is changed in the similarity judgment of the extraction model, the attribute "season" and the name "seasonal importance change" are associated in the table. In such a case, the name generating unit 17 generates a name of the extraction model such as "seasonal importance change model". In addition, the name generating unit 17 may generate a name of the extraction model by combining the names extracted from the table based on the attributes of the learning data and the feature quantities with high importance. For example, in a case where the season in which a product is sold and its design are important in determining the similarity of an extraction model, "season" indicating an attribute, "season-oriented" indicating a name, and "design" indicating an attribute, "design-oriented" indicating a name, are associated in the table. In such a case, the name generation unit 17 generates a name for the extraction model, such as "seasonal design-oriented model." In addition, the name generation unit 17 may generate a name for the extraction model by further combining the name extracted from the table with the name of the extraction model, for example, based on a change in the setting value.

[0053] The name generation unit 17 may generate the name of the extraction model using a name generation model. The name generation model is generated, for example, by deep learning using a neural network. The name generation model is generated, for example, by learning the relationship between a feature value that has a large influence on similarity judgment and the name of the extraction model. The name generation model may be generated by learning the relationship between the type of target product and a feature value that has a large influence on similarity judgment and the name of the extraction model. The learning data used to generate the name generation model is not limited to the above. The name generation model is generated, for example, in a system external to the extraction system 10. The name generation model may be generated by a learning means (not shown) in the extraction system 10.

[0054] FIG. 2 is a diagram showing an example of a screen displaying the extraction result of a product similar to the target product. In the example of the display screen in FIG. 2, three products similar to the target product are displayed. The output unit 13 outputs the screen shown in the example of the display screen in FIG. 2 based on the result extracted by the extraction unit 11 using the extraction model, for example. In the example of the display screen in FIG. 2, images of each product similar to the target product are displayed. In addition, in the example of the display screen in FIG. 2, data of "product name", "brand", "price", "color" and "release date" are displayed for each product similar to the target product. "Product name" is, for example, the name of the product. "Brand" is, for example, the name of the brand selling the product. "Price" is the selling price of the product. "Color" is, for example, the base color of the product. In addition, "release date" is, for example, the month in which the product was released. For example, the output unit 13 outputs a feature that has a higher contribution to the similarity judgment than other feature amounts among the feature amounts used by the extraction model for the similarity judgment to a screen showing the extraction result. It outputs a screen shown in the example of the display screen in FIG. 2. In addition, in the example of the display screen in FIG. 2, the similarity is displayed for each product similar to the target product. The output unit 13 outputs, for example, the product deviation similarity for the target product in the similarity judgment performed by the extraction model to a screen showing the extraction result. The screen shown in the example of the display screen in FIG. 2 is output. In addition, in the example of the display screen in FIG. 2, a "Change Settings" button is displayed. The "Change Settings" button is, for example, a button that displays a screen for changing the setting value of the feature importance. The screen for changing the setting value of the feature importance is, for example, a screen that displays the feature used by the extraction model for similarity judgment, the setting value of the feature importance, and a change field for changing the setting value of the feature importance. The output unit 13 outputs, for example, the "Change Settings" button shown in the example of the display screen in FIG. 2 to the screen showing the extraction result. In addition, when the button is pressed on the output screen, the acquisition unit 14 acquires information indicating that the button has been pressed. When the information indicating that the button has been pressed is acquired, the output unit 13 outputs a screen for changing the setting value of the feature importance.

[0055] FIG. 5 is a diagram showing an example of a display screen for changing the set value of the importance of a feature. In the example of the display screen in FIG. 5, "brand", "color", "release date" and "price" are displayed as feature values. In addition, in the example of the display screen in FIG. 5, the set value of the importance of each feature is displayed. In the example of the display screen in FIG. 5, "3" is displayed as the set value of the importance of each feature. In the example of the display screen in FIG. 5, "+" and "-" buttons are displayed for each feature. The "+" button is, for example, a button for increasing the set value of the feature by clicking on the screen. In addition, the "-" button is, for example, a button for decreasing the set value of the feature by clicking on the screen. The "+" and "-" buttons are examples of change columns for changing the set value of the importance of a feature. In addition, in the example of the display screen in FIG. 5, an "update" button is displayed. The "update" button is a button for reflecting the changed set value of the importance of the feature in the extraction of similar products by the extraction model.

[0056] The output unit 13 outputs, for example, a screen shown in the example of the display screen in FIG. 5. Then, when a change value is input in the change column on the output screen and the "Update" button is pressed, the acquisition unit 14 acquires the change value for changing the setting value of the feature importance and information indicating that the button has been pressed. The change value for changing the setting value of the feature importance and the update unit 15 updates the setting value of the feature importance in the extraction model, for example, based on the acquired change value. When the "Update" button is pressed, extraction of a product similar to the target product may be performed using the extraction model in which the setting value of the feature importance has been changed. In this case, when the change value for changing the setting value of the feature importance and information indicating that the button has been pressed are acquired, the update unit 15 updates the setting value of the feature importance in the extraction model, for example, based on the acquired change value. Then, the extraction unit 11 extracts a product similar to the target product, for example, based on the setting value of the feature importance after the update. When extraction of a product similar to the target product is performed using an extraction model that uses the changed feature importance setting value, the output unit 13 outputs, for example, a screen displaying the newly performed extraction results in a manner similar to the example display screen of Figure 2.

[0057] FIG. 6 is a diagram showing an example of a display screen for selecting learning data used for generating an extraction model or for relearning an extraction model. The learning data selected in the example of the display screen in FIG. 6 may be used for generating an extraction model. In the example of the display screen in FIG. 6, the name of the extraction model is displayed as "T model". In addition, in the example of the display screen in FIG. 6, images of each product similar to the target product are displayed. In addition, in the example of the display screen in FIG. 6, data of "product name", "brand", "price", "color" and "release date" are displayed for each product similar to the target product. The output unit 13 outputs a screen for selecting a product to be used as learning data in generating an extraction model or for relearning an extraction model, as shown in the example of the display screen in FIG. 6, for example.

[0058] In addition, in the example of the display screen of FIG. 6, a button for selecting a product to be used as learning data is displayed for each product. In addition, in the example of the display screen of FIG. 6, for example, by clicking on the screen, a product displayed in black is used as learning data. In the example of the display screen of FIG. 6, when a button displayed in black is clicked, it is displayed in white and the selection is cancelled. In addition, in the example of the display screen of FIG. 6, a button "Execute" is displayed. When the "Execute" button is pressed, machine learning using the selected learning data is executed. The machine learning using the selected learning data may include re-learning of the extraction model. When a product to be used as learning data is selected on the screen output by the output unit 13, the acquisition unit 14 acquires the selection result of the product to be used as learning data. In addition, when the "Execute" button is pressed on the screen, the acquisition unit 14 acquires, for example, information indicating that the button has been pressed. When information indicating that the "Execute" button has been pressed is acquired, the model update unit 16 uses the product selected in the selection result as learning data to re-learn the extraction model.

[0059] FIG. 7 is a diagram showing an example of a display screen for selecting learning data, in which a feature to be used in the learning data can be further selected. In the example of the display screen of FIG. 7, in addition to the display contents similar to those of the example of the display screen of FIG. 6, a check box for selecting a feature to be used in learning for each product is displayed as "feature selection". The output unit 13 outputs, for example, a screen for selecting a feature to be used in the learning data, as shown in the example of the display screen of FIG. 7. In the example of the display screen of FIG. 7, a feature with a check mark in the check box is used for re-learning. When the check box is clicked with a check mark in the check box, the selection is cancelled. In the example of the display screen of FIG. 7, an "execute" button is displayed. When the "execute" button is pressed, machine learning is performed using the selected learning data and feature. The feature selected in the example of the display screen of FIG. 7 may be used to generate an extraction model. When the "execute" button is pressed in a state in which a feature is selected on the screen output by the output unit 13, the acquisition unit 14 acquires, for example, a selection result of the feature to be used as learning data and information indicating that the "execute" button has been pressed. When the selection result of the features to be used as learning data and information indicating that the “Execute” button has been pressed are acquired, the model update unit 16 re-learns the extraction model using the features selected in the selection result as learning data.

[0060] The terminal device 20 is, for example, a terminal device used by a user to operate the extraction system 10. The terminal device 20 acquires, for example, from the output unit 13 of the extraction system 10, a screen displaying feature amounts used by the extraction model for similarity judgment, setting values ​​of the feature amount importance, and a change field for changing the setting value of the feature amount importance. Then, the terminal device 20 displays the acquired screen on a display device (not shown).

[0061] The terminal device 20 acquires, for example, a change value of the importance of a feature input by a user's operation on a screen output on a display device. The screen output by the terminal device 20 on the display device is a screen acquired by the terminal device 20 from the output unit 13 of the extraction system 10. Then, the terminal device 20 outputs the change value of the importance of a feature to the acquisition unit 14 of the extraction system 10.

[0062] The terminal device 20 outputs, for example, a screen for selecting learning data to be used for generating an extraction model or for relearning an extraction model from the output unit 13 of the extraction system 10. Then, the terminal device 20 displays the screen acquired from the output unit 13 of the extraction system 10 on a display device (not shown). Furthermore, image data of the screen may be generated in the terminal device 20. When image data of the screen is generated in the terminal device 20, the terminal device 20 acquires an instruction for the display content from the output unit 13 of the extraction system 10.

[0063] The terminal device 20 acquires the selection result of the learning data to be used for generating the extraction model or for re-learning the extraction model, which is input by a user's operation on a screen displayed on the display device, for example. The input may include an operation of the user selecting from among the candidates displayed on the screen. Then, the terminal device 20 outputs the selection result of the learning data to be used for generating the extraction model or for re-learning the extraction model to the acquisition unit 14 of the extraction system 10.

[0064] For example, a personal computer, a tablet computer, or a smartphone can be used as the terminal device 20. In addition, the terminal device 20 may be an integrated system with the extraction system 10.

[0065] The data management device 30 stores the product name and product attribute data. The data management device 30 may further store a photographed image of the product. There may be a plurality of data management devices 30. When there are a plurality of data management devices 30, each data management device 30 may store the product name and product attribute data related to a different industry.

[0066] The following describes a process for changing the setting value of the importance of a feature amount in an extraction model in the extraction system 10. Fig. 8 is a diagram showing an example of an operational flow in a process for changing the setting value of the importance of a feature amount.

[0067] The output unit 13 outputs a screen displaying the feature amounts used by the extraction model for similarity judgment, the setting values ​​of the feature amount importance, and a change field for changing the setting value of the feature amount importance (step S11). The extraction model is a learning model that extracts products similar to the target product. The output unit 13 outputs, for example, to the terminal device 20, a screen displaying the feature amounts used by the extraction model for extracting products similar to the target product for similarity judgment, the setting values ​​of the feature amount importance, and a change field for changing the setting value of the feature amount importance.

[0068] The terminal device 20 displays, for example, on a display device not shown, a screen showing features used by an extraction model that extracts products similar to the target product for similarity judgment, setting values ​​for the importance of the features, and a change field for changing the setting value of the importance of the features.

[0069] When a user operates the terminal device 20 to input a change value for changing the setting value of the importance of a feature on the screen displayed on the display device, the terminal device 20 outputs the input change value to the extraction system 10.

[0070] The acquisition unit 14 of the extraction system 10 acquires, for example, a change value for changing the setting value of the importance of the feature amount (step S12). The extraction system 10 acquires, for example, a change value for changing the setting value of the importance of the feature amount output from the terminal device 20.

[0071] When the change value for changing the setting value of the feature importance is acquired, the update unit 15 updates the setting value of the feature importance in the extraction model based on the change value for changing the setting value of the feature importance acquired by the acquisition unit 14 (step S13).

[0072] When the setting value of the importance of the feature amount in the extraction model is updated, the update unit 15 stores, for example, the updated extraction model (step S14). The update unit 15 stores, for example, the updated extraction model in the storage unit 12.

[0073] A description will be given of a process for extracting a product similar to a target product in the extraction system 10. Fig. 9 is a diagram showing an example of an operation flow in the process for extracting a product similar to a target product.

[0074] The acquisition unit 14 acquires, for example, information related to the target product (step S21). The acquisition unit 14 acquires, for example, attribute data of the target product as information related to the target product.

[0075] When the information on the target product is acquired, the extraction unit 11 extracts products similar to the target product, for example, by using an extraction model (step S22). The extraction unit 11 extracts products similar to the target product, for example, by using the extraction model based on the information on the target product.

[0076] When a product similar to the target product is extracted, the output unit 13 outputs, for example, a screen displaying the extraction results of products similar to the target product (step 23). The output unit 13 also outputs, for example, a screen displaying the extraction results together with a button for selecting an operation to change the setting value of the importance of the feature. The output unit 13 outputs, for example, to the terminal device 20, a screen displaying the extraction results of products similar to the target product. The terminal device 20 displays, for example, a screen displaying the extraction results of products similar to the target product on a display device not shown.

[0077] When a button for changing the setting value of the feature importance on the output screen is pressed by the user's operation to select changing the setting value of the feature importance (Yes in step S24), the output unit 13 outputs a screen displaying the feature used by the extraction model for similarity judgment, the setting value of the feature importance, and a change field for changing the setting value of the feature importance (step S25).The output unit 13 outputs to the terminal device 20, for example, a screen displaying the feature used by the extraction model for extracting products similar to the target product for similarity judgment, the setting value of the feature importance, and a change field for changing the setting value of the feature importance.

[0078] The terminal device 20 outputs, for example, on a display device (not shown) a screen displaying features used by an extraction model that extracts products similar to the target product for similarity judgment, a setting value for the importance of the feature, and a change field for changing the setting value of the importance of the feature.

[0079] When the user operates the terminal device 20 to change the importance of the feature amount, the terminal device 20 outputs the changed setting value of the importance of the feature amount to the extraction system 10 .

[0080] The extraction system 10 acquires the changed setting value of the importance of the feature amount from, for example, the terminal device 20 (step S26).

[0081] When the changed feature importance setting value is acquired, the update unit 15 updates the feature importance setting value in the extraction model based on the changed feature importance setting value acquired by the acquisition unit 14 (step S27).

[0082] When the setting value of the feature amount importance is updated, the extraction unit 11 extracts products similar to the target product, for example, by using the extraction model with the updated setting value (step S28).

[0083] When the product similar to the target product is extracted, the process returns to step S23, and the output unit 13 outputs, for example, the extraction result of the product similar to the target product.

[0084] Furthermore, in step S24, if changing the setting value of the feature importance is not selected (No in step S24), the extraction system 10 waits, for example, until a new operation is performed. Also, for example, an "OK" or "End" button may be displayed on the screen displaying the extraction result, and when the button is pressed on the screen, the extraction system 10 may end the process of extracting products similar to the target product.

[0085] A description will be given of a process for updating an extraction model by relearning in the extraction system 10. Fig. 10 is a diagram showing an example of an operational flow in a process for updating an extraction model by relearning.

[0086] The output unit 13 outputs, for example, a list of products that are candidates for learning data (step S31). The output unit 13 outputs, for example, to the terminal device 20, the list of products that are candidates for learning data.

[0087] The terminal device 20 outputs a list of products that are candidates for learning data to, for example, a display device (not shown). The terminal device 20 acquires a selection of products to be used as learning data input by a user's operation. The terminal device 20 outputs the selection of products to be used as learning data to, for example, the extraction system 10.

[0088] The acquisition unit 14 acquires a selection of products to be used as learning data (step S32).

[0089] When the selection of the product to be used as the learning data is acquired, the model update unit 16 re-learns the extraction model using the selected product and the attribute data of the selected product as the learning data (step S33).

[0090] When the extraction model is re-learned, the model update unit 16 updates the extraction model stored in the storage unit 12, for example, with the re-learned extraction model (step S34).

[0091] In addition, the extraction system 10 is applicable, for example, to the medical health care field. For example, the extraction system 10 can be used to extract products similar to a pre-release health food that is used to forecast the demand for the pre-release health food. For example, a person planning a new health food product can estimate the demand for the health food to be released based on products similar to the pre-release health food. In this way, by extracting products similar to the new health food product, it is possible to support the decision-making of the person planning the new health food product. Also. The extraction system 10 can be used, for example, to extract medicines similar to a new drug. The extraction system 10 may also extract people instead of products. For example, when a doctor decides on a medical treatment policy, the extraction system 10 extracts patients who have similar treatment histories and attributes to the patient who is the subject of the medical treatment. Then, the doctor can decide the medical treatment policy by referring to information on patients who have similar treatment histories and attributes to the patient who is the subject of the medical treatment. In addition, a nurse and a physical therapist may use the extraction system 10 to extract data on patients who have similar treatment histories and attributes and decide a treatment policy for the patient. In this manner, the extraction system 10 can be used to assist, for example, medical personnel in making decisions.

[0092] The extraction system 10 outputs a screen displaying feature quantities used by an extraction model that extracts products similar to the target product for similarity judgment, the feature quantity importance setting value, and a change field for changing the feature quantity importance setting value. The extraction system 10 also acquires the feature quantity importance change value input on the screen. The extraction system 10 then updates the feature quantity importance setting value in the extraction model based on the change value. In this way, by updating the feature quantity importance setting value in the extraction model based on the feature quantity importance change value input on the screen, for example, a user can easily adjust the extraction results of products similar to the target product. Therefore, by using the extraction system 10, it is possible to easily extract products similar to the target product.

[0093] In addition, the extraction system 10 further displays a screen that displays the extraction results obtained by extracting products similar to the target product using an extraction model after updating the setting value of the importance of the feature. By displaying the extraction results obtained by extracting products similar to the target product using the updated extraction model, the user can understand how the extraction results have changed due to, for example, changing the setting value of the importance of the feature. This makes it possible to more appropriately extract products similar to the target product.

[0094] In addition, the extraction system 10 outputs, for example, a screen on which a product to be used for re-learning the extraction model can be selected. By outputting a screen on which a product to be used for re-learning the extraction model can be selected, the user can easily select learning data to be used for re-learning the extraction model. Therefore, by using the extraction system 10, a product more suitable for the user's purpose can be extracted as a product similar to the target product.

[0095] Furthermore, the extraction system 10 outputs, for example, a screen on which the feature quantity to be used for re-learning the extraction model can be selected. By outputting a screen on which the feature quantity to be used for re-learning the extraction model can be selected, the user can easily select the feature quantity to be used for re-learning the extraction model. Therefore, by using the extraction system 10, a product more suitable for the user's purpose can be extracted as a product similar to the target product.

[0096] Furthermore, when an extraction model capable of explaining the reason for the extraction result is used, the extraction system 10 outputs, for example, a screen displaying the reason for the extraction result together with the extraction result. By outputting the reason together with the extraction result, the user can easily understand the extraction result. Furthermore, by outputting the reason together with the extraction result, the user can more appropriately select the feature to be changed and determine the change value when changing the importance of the feature amount.

[0097] In addition, the extraction system 10 generates the name of the extraction model based on, for example, the attributes of the learning data and the features with high importance. Therefore, when selecting an extraction model to be used for extraction from a plurality of extraction models, the user can easily select an extraction model suitable for the type or use of the target product based on the name. Therefore, by using the extraction system 10, it is possible to easily extract products suitable for the user's use.

[0098] Each process in the extraction system 10 may be distributed and executed in a plurality of information processing systems connected via a network. For example, the processes in the output unit 13, the acquisition unit 14, the update unit 15, and the extraction unit 11, and the processes in the model update unit 16 and the model update unit 16 may be executed in another information processing system. Also, for example, the processes in the output unit 13, the acquisition unit 14, the update unit 15, the extraction unit 11, and the model update unit 16, and the processes in the model update unit 16 may be executed in another information processing system. It can be appropriately set which of the plurality of information processing systems each process in the extraction system 10 is executed in. Also, some or all of the processes in the extraction system 10 may be executed in the terminal device 20.

[0099] Each process in the extraction system 10 can be realized by executing a computer program on a computer. Fig. 11 shows an example of the configuration of a computer 100 that executes a computer program that performs each process in the extraction system 10. The computer 100 includes a CPU (Central Processing Unit) 101, a memory 102, a storage device 103, an input / output I / F (Interface) 104, and a communication I / F 105.

[0100] The CPU 101 reads out and executes a computer program for performing each process from the storage device 103. The CPU 101 may be configured by a combination of a plurality of CPUs. The CPU 101 may also be configured by a combination of a CPU and another type of processor. For example, the CPU 101 may be configured by a combination of a CPU and a GPU (Graphics Processing Unit). The memory 102 is configured by a DRAM (Dynamic Random Access Memory) or the like, and temporarily stores the computer program executed by the CPU 101 and data being processed. The storage device 103 stores the computer program executed by the CPU 101. The storage device 103 is configured by, for example, a non-volatile semiconductor storage device. Other storage devices such as a hard disk drive may be used for the storage device 103. The input / output I / F 104 is an interface that accepts input from an operator and outputs display data and the like. The communication I / F 105 is an interface that transmits and receives data between the terminal device 20 and the data management device 30. The terminal device 20 and the data management device 30 may also be configured in the same manner.

[0101] The computer programs used to execute each process can also be distributed by storing them in a computer-readable recording medium that non-temporarily records data. As the recording medium, for example, a magnetic disk such as a magnetic tape for recording data or a hard disk can be used. Also, an optical disk such as a CD-ROM (Compact Disc Read Only Memory) can be used as the recording medium. A non-volatile semiconductor storage device can also be used as the recording medium.

[0102] A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes.

[0103] [Appendix 1] an output means for outputting a screen displaying feature values ​​used in a similarity judgment by an extraction model for extracting products similar to the target product, a setting value of the importance of the feature values, and a change field for changing the setting value of the importance of the feature values; an acquisition means for acquiring a change value for changing a setting value of the importance of the feature amount, the change value being input into the change field of the output screen; an update means for updating a setting value of the importance of the feature amount in the extraction model based on the changed value; An extraction system comprising:

[0104] [Appendix 2] Further comprising an extraction means for extracting products similar to the target product using the extraction model; the output means outputs, together with the extraction result, a feature used by the extraction model for the similarity determination, a setting value of the importance of the feature, and a change field for changing the setting value. 2. The extraction system according to claim 1.

[0105] [Appendix 3] The extraction means re-extracts products similar to the target product using the extraction model based on the updated importance setting value. 2. An extraction system as described in appended claim 2.

[0106] [Appendix 4] The output means outputs a screen for selecting a product to be used as learning data from the products extracted by the extraction model; the acquiring means acquires a selection result of a product to be used as learning data, selected on the output screen, Further comprising a model updating means for updating the extraction model by performing machine learning using the product selected in the selection result as learning data. 4. The extraction system according to claim 2 or 3.

[0107] [Appendix 5] the output means outputs a screen for selecting a feature quantity to be used as learning data from the feature quantities used in learning the extraction model; the acquiring means acquires a selection result of the feature quantity to be used as the learning data, selected on the output screen; a model updating means for updating the extraction model by executing machine learning using the feature quantity selected in the selection result as learning data; 4. The extraction system according to claim 2 or 3.

[0108] [Appendix 6] the output means outputs a screen further displaying a feature amount that contributed to the similarity determination with the target product when outputting a screen displaying the product extracted by the extraction model. 4. The extraction system according to claim 2 or 3.

[0109] [Appendix 7] the output means outputs a screen for selecting one of a plurality of extraction models; The acquisition means acquires a selection result of an extraction model selected on the output screen, the output means outputs a screen displaying a feature amount in the extraction model selected in the selection result, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount. 2. The extraction system according to claim 1.

[0110] [Appendix 8] a name generating means for generating a name of the extraction model based on at least one of a feature value used as learning data for the extraction model and an importance of the feature value; 2. The extraction system according to claim 1.

[0111] [Appendix 9] the output means outputs a screen including a button for switching between a screen displaying a product extracted by the extraction model and a screen displaying a change field for changing a setting value of the importance of the feature amount in the extraction model. 2. The extraction system according to claim 1.

[0112] [Appendix 10] the output means outputs a screen displaying the feature amounts that contributed to the similarity judgment by the extraction model in order of the contribution to the similarity judgment. 7. The extraction system according to claim 6.

[0113] [Appendix 11] the output means outputs, as a change field for changing the set value of the importance of the feature, buttons for increasing or decreasing the set value by operating the buttons on the screen. 2. The extraction system according to claim 1.

[0114] [Appendix 12] The output means further outputs a standard feature value and an importance of the feature value in the industry in which the target product is distributed. 2. The extraction system according to claim 1.

[0115] [Appendix 13] outputting a screen displaying feature amounts used in a similarity judgment by an extraction model that extracts products similar to the target product, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount; obtaining a change value for changing the setting value of the importance of the feature amount, which is to be input into the change field of the output screen; updating a setting value of the importance of the feature amount in the extraction model based on the changed value; Extraction method.

[0116] [Appendix 14] Extracting products similar to the target product using the extraction model; outputting, together with the extraction result, the feature amount used by the extraction model for the similarity judgment, a setting value of the importance of the feature amount, and a change field for changing the setting value; The extraction method described in Appendix 13.

[0117] [Appendix 15] re-extracting products similar to the target product using the extraction model based on the updated importance setting value; The extraction method described in Appendix 14.

[0118] [Appendix 16] outputting a screen for selecting a product to be used as learning data from the products extracted by the extraction model; Acquire the selection results of the products to be used as learning data, selected on the output screen; updating the extraction model by performing machine learning using the products selected in the selection result as learning data; 16. The extraction method according to claim 14 or 15.

[0119] [Appendix 17] outputting a screen for selecting a feature to be used as learning data from the feature used in learning the extraction model; acquiring a selection result of the feature quantity to be used as the learning data, selected on the output screen; updating the extraction model by performing machine learning using the feature quantity selected in the selection result as learning data; 16. The extraction method according to claim 14 or 15.

[0120] [Appendix 18] When outputting a screen displaying the product extracted by the extraction model, a screen displaying the feature amount that contributed to the similarity judgment with the target product is further output. 16. The extraction method according to claim 14 or 15.

[0121] [Appendix 19] A screen is output for selecting one of the extraction models from multiple extraction models, Acquire a selection result of an extraction model selected on the output screen, outputting a screen displaying the feature amount in the extraction model selected in the selection result, the setting value of the feature amount importance, and a change field for changing the setting value of the feature amount importance; The extraction method described in Appendix 13.

[0122] [Appendix 20] generating a name for the extraction model based on at least one of a feature amount used as learning data for the extraction model and an importance of the feature amount; The extraction method described in Appendix 13.

[0123] [Appendix 21] outputting a screen including a button for switching between a screen displaying a product extracted by the extraction model and a screen displaying a change field for changing a setting value of the importance of the feature amount in the extraction model; The extraction method described in Appendix 13.

[0124] [Appendix 22] outputting a screen displaying the feature amounts that contributed to the similarity judgment by the extraction model in order of the contribution to the similarity judgment; The extraction method described in Appendix 18.

[0125] [Appendix 23] the output means outputs, as a change field for changing the set value of the importance of the feature, buttons for increasing or decreasing the set value by operating the buttons on the screen. The extraction method described in Appendix 13.

[0126] [Appendix 24] Further outputting standard feature values ​​and feature importance in the industry in which the target product is distributed; The extraction method described in Appendix 13.

[0127] [Appendix 25] A process of outputting a screen displaying feature amounts used in a similarity judgment by an extraction model that extracts products similar to the target product, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount; A process of acquiring a change value for changing the setting value of the importance of the feature amount, which is input into the change field of the output screen; updating a setting value of the importance of the feature in the extraction model based on the changed value; An extraction program that causes a computer to execute the above.

[0128] [Appendix 26] A process of extracting products similar to the target product using the extraction model; a process of outputting, together with the extraction result, the feature amount used by the extraction model for the similarity judgment, a setting value of the importance of the feature amount, and a change field for changing the setting value; 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0129] [Appendix 27] A process of re-extracting products similar to the target product using the extraction model based on the updated importance setting value. 27. The extraction program according to claim 26, which causes a computer to execute the above steps.

[0130] [Appendix 28] A process of outputting a screen for selecting a product to be used as learning data from the products extracted by the extraction model; A process of acquiring a selection result of a product to be used as learning data, which is selected on the output screen; A process of updating the extraction model by performing machine learning using the products selected in the selection result as learning data; 28. The extraction program according to claim 26 or 27, which causes a computer to execute the above.

[0131] [Appendix 29] outputting a screen for selecting a feature to be used as learning data from the feature used in learning the extraction model; A process of acquiring a selection result of a feature quantity to be used as the learning data, selected on the output screen; and updating the extraction model by performing machine learning using the features selected in the selection result as learning data.

[0132] [Appendix 30] A process of outputting a screen that further displays the features that contributed to the similarity judgment with the target product when outputting a screen that displays the product extracted by the extraction model. 28. The extraction program according to claim 26 or 27, which causes a computer to execute the above.

[0133] [Appendix 31] A process of outputting a screen for selecting one of the plurality of extraction models; A process of acquiring a selection result of an extraction model selected on the output screen; A process of outputting a screen displaying the feature amount in the extraction model selected in the selection result, the setting value of the feature amount importance, and a change field for changing the setting value of the feature amount importance. 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0134] [Appendix 32] A process of generating a name of the extraction model based on at least one of a feature amount used as learning data for the extraction model and an importance of the feature amount. 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0135] [Appendix 33] A process to output a screen containing a button to switch between a screen that displays a change field for changing the setting value of the feature importance in the 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0136] [Appendix 34] A process of outputting a screen that displays the feature amounts that contributed to the similarity judgment by the extraction model in order of their contribution to the similarity judgment. The extraction program according to claim 30, which causes a computer to execute the above steps.

[0137] [Appendix 35] A process of outputting an image that can be operated on the screen to increase or decrease the setting value as a change field for changing the setting value of the importance of the feature. 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0138] [Appendix 36] Further output of standard feature values ​​and feature importance in the industry in which the target product is distributed 26. The extraction program according to claim 25, which causes a computer to execute the above steps.

[0139] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-mentioned embodiments. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be appropriately combined with other embodiments. [Explanation of symbols]

[0140] 10 Extraction System 11 Extraction part 12 Storage section 13 Output section 14 Acquisition Department 15 Update section 16 Model Update Section 17 Name generator 20 Terminal Equipment 30 Data management device 100 Computers 101 CPU 102 Memory 103 Storage device 104 Input / Output Interface 105 Communication I / F

Claims

1. an output means for outputting a screen displaying feature values ​​used in a similarity judgment by an extraction model for extracting products similar to the target product, a setting value of the importance of the feature values, and a change field for changing the setting value of the importance of the feature values; an acquisition means for acquiring a change value for changing a setting value of the importance of the feature amount, the change value being input to the change field of the output screen; an update means for updating a setting value of the importance of the feature amount in the extraction model based on the changed value; An extraction system comprising:

2. Further comprising an extraction means for extracting products similar to the target product using the extraction model; the output means outputs, together with the extraction result, a feature used by the extraction model for the similarity determination, a setting value of the importance of the feature, and a change field for changing the setting value. The extraction system of claim 1 .

3. the extraction means re-extracts products similar to the target product using the extraction model based on the updated setting value of the importance of the feature amount; The extraction system of claim 2.

4. The output means outputs a screen for selecting a product to be used as learning data from the products extracted by the extraction model; the acquiring means acquires a selection result of a product to be used as learning data, selected on the output screen, Further comprising a model updating means for updating the extraction model by performing machine learning using the product selected in the selection result as learning data. An extraction system according to claim 2 or 3.

5. the output means outputs a screen further displaying a feature amount that contributed to the similarity determination with the target product when outputting a screen displaying the product extracted by the extraction model. An extraction system according to claim 2 or 3.

6. the output means outputs a screen for selecting one of a plurality of extraction models; The acquisition means acquires a selection result of an extraction model selected on the output screen, the output means outputs a screen displaying a feature amount in the extraction model selected in the selection result, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount. The extraction system of claim 1 .

7. a name generating unit that generates a name of the extraction model based on at least one of a feature value and an importance of the feature value used as learning data for the extraction model; The extraction system of claim 1 .

8. the output means outputs a screen including a button for switching between a screen displaying a product extracted by the extraction model and a screen displaying a change field for changing a setting value of the importance of the feature amount in the extraction model. The extraction system of claim 1 .

9. outputting a screen displaying feature amounts used in a similarity judgment by an extraction model that extracts products similar to the target product, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount; obtaining a change value for changing the setting value of the importance of the feature amount, which is to be input into the change field of the output screen; updating a setting value of the importance of the feature amount in the extraction model based on the changed value; Extraction method.

10. A process of outputting a screen displaying feature amounts used in a similarity judgment by an extraction model that extracts products similar to the target product, a setting value of the importance of the feature amount, and a change field for changing the setting value of the importance of the feature amount; A process of acquiring a change value for changing the setting value of the importance of the feature amount, which is input into the change field of the output screen; updating a setting value of the importance of the feature in the extraction model based on the changed value; An extraction program that causes a computer to execute the above.

Images