JP2018128806A - Extraction apparatus, extraction method, and extraction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To flexibly show the characteristics of a commercial material.SOLUTION: An extraction apparatus includes an acquisition unit and an extraction unit. The acquisition unit acquires commercial material information associated with a commercial material to be transacted on a network. The extraction unit extracts characteristics information which is information satisfying a predetermined condition in a group, of commercial material information of a predetermined commercial material, in the predetermined commercial material included in the group formed based on the contents designated by a user. For example, the extraction unit extracts the commercial material information, as characteristics information, when the commercial material information of the predetermined commercial material included in the group is information showing superiority in the group.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an extraction apparatus, an extraction method, and an extraction program.

  In recent years, with the rapid spread of the Internet, various contents are actively provided through a network. For example, when the content is a product, the user performs a search by specifying a category of the product, and accesses an introduction page or a sales page of the content displayed as a search result.

  Here, as an example of the technology using the search result related to the content, the ranking of the target product is generated from the sales history of the target product group searched by the user, and the generated ranking is used as the text of the advertising content. Technology is known.

JP-A-2015-79449

  However, in the above-described conventional technology, it is not always possible to appeal the features relating to the merchandise more flexibly. Specifically, in the above-described conventional technology, the text of the advertising content is automatically generated from features other than the features to be ranked, and in a page that actually introduces a product (such as a web page) It does not present the features of the product to the user. For this reason, there may be a case where various features of the commercial material cannot be appealed to the user.

  The present application has been made in view of the above, and an object of the present invention is to provide an extraction device, an extraction method, and an extraction program that can more flexibly appeal the features related to commercial products.

  The extraction apparatus according to the present application includes an acquisition unit that acquires product information associated with a product that is a target of a commercial transaction on a network, and a predetermined group included in a set that is formed based on content specified by a user. The product includes an extraction unit that extracts, from the product information of the predetermined product, feature information that is information satisfying a predetermined condition in the set.

  According to one aspect of the embodiment, there is an effect that it is possible to appeal the features relating to the product more flexibly.

FIG. 1 is a diagram illustrating an example of an extraction process according to the embodiment. FIG. 2 is a diagram illustrating a configuration example of the extraction processing system according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of the extraction device according to the embodiment. FIG. 4 is a diagram illustrating an example of the product information storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of a learning information storage unit according to the embodiment. FIG. 6 is a diagram illustrating an example of the provided information storage unit according to the embodiment. FIG. 7 is a diagram illustrating a configuration example of the web server according to the embodiment. FIG. 8 is a flowchart (1) illustrating a processing procedure according to the embodiment. FIG. 9 is a flowchart (2) illustrating a processing procedure according to the embodiment. FIG. 10 is a flowchart (3) illustrating the processing procedure according to the embodiment. FIG. 11 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the extraction device.

  Hereinafter, a mode for carrying out an extraction apparatus, an extraction method, and an extraction program according to the present application (hereinafter referred to as “embodiment”) will be described in detail with reference to the drawings. Note that the extraction apparatus, the extraction method, and the extraction program according to the present application are not limited to the embodiment. In addition, the embodiments can be appropriately combined within a range that does not contradict processing contents. In the following embodiments, the same portions are denoted by the same reference numerals, and redundant description is omitted.

[1. Example of extraction process)
First, an example of the extraction process according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of an extraction process according to the embodiment. FIG. 1 shows an example in which a page (content) introducing a product in electronic commerce is extracted by the extraction apparatus 100 according to the present application.

  An extraction apparatus 100 illustrated in FIG. 1 is a server apparatus that performs a process of generating a page (an example of content) that introduces a product in electronic commerce. Specifically, the extraction apparatus 100 extracts information that characterizes a product based on a web page related to the product or information associated with the product. Then, the extraction apparatus 100 adds a page that can appeal the features of the product to the user by adding the extracted features to the original web page (hereinafter sometimes referred to as “appeal page”). Generate. In FIG. 1, a real estate rental property will be described as an example of a commercial product. In FIG. 1, a page (web page) is shown as an example of content. However, the content is not limited to a page, and includes various information content such as advertisement content and an information page provided through an application.

  A user terminal 10 shown in FIG. 1 is an information processing terminal used by a user. For example, the user terminal 10 is a smartphone or a tablet terminal. The user terminal 10 uses a service on the network in accordance with a user operation. For example, the user terminal 10 can access an electronic commerce site (hereinafter referred to as a “commerce site”) and make a purchase application for various services or make an inquiry. In the embodiment, the user terminal 10 may be read as a user. For example, the description “the user accesses the website” may actually indicate a situation where “the user terminal 10 operated by the user accesses the website”.

  Here, in a commercial transaction site such as the real estate information site exemplified in the embodiment, a plurality of business operators (for example, real estate agents) upload various products (the property in the example of FIG. 1) to each site. It shall be.

  In such a commercial transaction site, a property title and an explanation are usually created by a business operator and associated with property information (commodity information). In addition, companies also provide detailed information on the property, such as the rent of the property, the nearest station, the distance to the nearest station, the area (occupied area), the age of the building, and whether pets are allowed. Correspond as the product information of the property.

  Then, the business operator uploads an image of the property, other product information, and the like along with the property title and description. The commerce site side generates a property introduction page based on information uploaded from the business. A user who uses a commercial transaction site accesses the commercial transaction site and refers to the title and explanation posted on the introduction page to make an application or inquiry to the property.

  Here, in such a commercial transaction site, creation of information to be given to the property (for example, description of the property) may be entrusted to the business operator. For this reason, in the commercial transaction site, there is a problem that appropriate property information is not always provided to the user. In other words, when a business operator arbitrarily gives a description, for example, an SEO measure (Search Engine Optimization) is applied, so that an excessively long description is given or unrelated words are inserted into the description. Sometimes.

  In this case, when the user accesses the commercial transaction site and browses the introduction page, it may be difficult to determine what characteristics the property has. Thereby, the usability of the whole commerce site may be reduced.

  In addition, when the content described on the property introduction page is entrusted to the business operator, the characteristics of the property may not be fully appealed. For example, the property accessed by the user is a property with a relatively low rent or a pet can be bred in a set designated by the user as a search target (for example, a property group having a common nearest station). Suppose that there is an advantage. However, since it is difficult for business operators to determine whether such information can be an appealing point compared to other properties, there is no particular emphasis on the introduction page (for example, description of the property). There is a case. In addition, even if the business operator has described the characteristics, the user who browses the property page may not be aware of it.

  Here, it is possible to obligate businesses to give appropriate explanations at the request of the commerce site, but for businesses that upload a large number of properties, etc. Giving an appropriate description is a heavy workload. For this reason, a business operator may hesitate to upload to the commercial transaction site. In this case, the commerce site side loses a large number of operators, which is not desirable for the operation of the commerce site. In addition, the number of uploads is enormous because the commercial site gives artificially appropriate explanations (for example, explanations clearly showing appeal points) to all uploaded properties. In some cases it is not realistic.

  Therefore, the extraction apparatus 100 according to the embodiment extracts information indicating the characteristics of the property from the product information associated with the property by the business operator by a method described below. Further, the extraction apparatus 100 generates an appeal page that is a page that introduces a property and is provided with feature information (appeal point) for promoting the property based on the extracted feature. With this process, the extraction device 100 can generate a page that appeals the characteristics of the property to the user without imposing a burden on the business operator or the commercial transaction site. Thereby, since the extraction apparatus 100 can appeal the characteristic regarding a commercial material more flexibly, it can improve the appeal effect with respect to a user. Hereinafter, an example of the extraction process performed by the extraction apparatus 100 will be described along the flow with reference to FIG.

  As illustrated in FIG. 1, the extraction apparatus 100 performs learning for performing processing for extracting feature information from product information as a learning phase. First, the extraction device 100 acquires product information as learning data to be used for the learning process (step S11).

  Specifically, the extraction apparatus 100 acquires the commercial materials B01, B02, B03 and the like that are existing property introduction pages uploaded to the commercial transaction site. The product B01 or the like is, for example, information on a property uploaded by a business operator, and includes product information that is detailed information indicating specifications of each property. The learning data may be an introduction page such as the product B01 generated based on information uploaded by the business operator.

  For example, the extraction apparatus 100 acquires information on the product B01 and the like by crawling on the network using a known method. Alternatively, the extraction device 100 acquires information such as the product B01 from a predetermined web server that provides the real estate information site. Note that the extraction apparatus 100 does not necessarily have to acquire the product B01 or the like from the same commercial transaction site. That is, the extraction apparatus 100 may acquire learning data from any commercial transaction site as long as the property and the product information associated with the property can be acquired.

  Subsequently, the extraction apparatus 100 creates a database of items that can be characteristics of the product information based on the acquired product information (step S12). For example, when the product is real estate such as a property, the extraction apparatus 100 permits the rent of the property, the nearest station, the distance to the nearest station, the size (occupied area), the age of building, and pet breeding. Whether or not to do so is determined as an item that can be a feature of the product information. Note that the extraction apparatus 100 may accept settings of items that can be artificially featured in advance for the determination process. For example, the extraction apparatus 100 accepts in advance settings of items that can be characteristic for each category of goods (for example, real estate). With this process, for example, when the product to be processed is acquired, the extraction apparatus 100 can quickly extract information corresponding to items that can be characteristic information in the product information.

  In addition, the extraction device 100 learns about a portion that can be characteristic information (appeal point) of the commercial product in the explanation (text data) associated with the commercial product in the acquired commercial product information (step S13). ). That is, the extraction device 100 generates a predetermined model (learning device) for extracting features included in the text data.

  Although details will be described later, the extraction apparatus 100 is not limited to one method, and may generate a model using various methods. For example, the extraction apparatus 100 uses, as learning data, a summary sentence obtained by manually summarizing the explanation sentence together with the explanation sentence. That is, the extraction apparatus 100 performs machine learning using a summary sentence obtained by artificially summarizing an explanatory sentence as correct answer data, and using a phrase extracted in the process as feature information in the explanatory sentence. In this case, the extraction apparatus 100 can learn the features of the words and phrases extracted during manual summarization processing by using sufficient samples for learning. For this reason, when a new explanatory note is input to the model, the extraction apparatus 100 can extract a word that is assumed to be important from the explanatory note, imitating a process of manually creating a summary sentence. This extracted information can be characteristic information of an explanatory text (in other words, a property associated with the explanatory text).

  Further, the extraction device 100 may generate a model by learning through statistical processing. For example, the extraction apparatus 100 performs learning for extracting words that are included in a certain explanatory text and that characterize the explanatory text. Such processing is performed, for example, by calculating the importance of each word in the document when the explanatory text is regarded as one document, and generating a rule for extracting a phrase from the document based on the importance. . For example, when an explanatory text (document) is input based on an index value such as tf-idf (Term Frequency-Inverse Document Frequency), the extraction apparatus 100 may characterize the explanatory text in the explanatory text. A model that extracts (outputs) may be generated.

  Further, the extraction apparatus 100 may generate a model using, for example, a deep learning technique. That is, the extraction apparatus 100 performs learning for extracting a portion that can be a feature in the explanatory text using a deep learning technique. For example, the model generated by the extraction apparatus 100 is a model in which a plurality of nodes that output calculation results for input data are connected in multiple layers, and is a model for learning features included in text data. For example, the model is a neural network in which layers having a plurality of nodes are connected in multiple stages, and is a DNN (Deep Neural Network) realized by a so-called deep learning technique.

  For example, the model is generated by the following learning method using the deep learning technique. In the model, the connection coefficient between each node is initialized, and text having various words is input. Then, the model is generated by processing such as backpropagation (Backpropagation (error back propagation method)) for correcting a parameter (connection coefficient) so that an error between an output in the model and an input is reduced. For example, the model is learned so that, when text data constituting the explanatory text is input, characteristic information (for example, a phrase characterizing the explanatory text) included in the explanatory text is output. In the learning process, adjustment such as returning an appropriate feedback manually may be performed on the output of the model.

  Note that the model learning method and the generated model are not limited to the method described above, and any known technique can be applied. In other words, the extraction device 100 can use any model as long as it can output feature information indicating features abstracted from text data. The extraction device 100 stores the model generated in the learning phase in a predetermined storage unit. The above is the learning phase according to the embodiment.

  Thereafter, when the user uses a real estate information site or the like as the generation phase, the extraction apparatus 100 extracts points that are particularly appealing points from the explanatory text associated with the property, and based on the extracted information Process to generate appeal page.

  For example, the user designates a product (property) that he / she wants to browse by inputting conditions related to the product on the real estate information site (step S21). For example, the user narrows down the properties desired to be browsed by specifying conditions such as the area where the property is located and the nearest station. And a user shall browse the predetermined | prescribed property XXX among the narrowed properties in the example of FIG. For example, it is assumed that the user clicks a link for browsing a predetermined property XXX from the property group presented after specifying the conditions.

  At this time, the extraction apparatus 100 specifies a set formed based on the content specified by the user. Here, the set refers to a plurality of properties that match the content specified by the user among the properties that the extraction apparatus 100 has created in the database. For example, the set is common to the property group narrowed down by the conditions specified by the user.

  Then, the extraction apparatus 100 extracts, as feature information, items that are superior to the property included in the set among the product information included in the property that the user intends to browse (step S22). For example, the extraction device 100 extracts items that are superior when compared with the product information of a plurality of properties included in the set from the product information associated with the product XXX.

  For example, the extraction apparatus 100 refers to a database that has been built in advance, and among the product information associated with the property XXX, the rent is relatively low compared to the set (for example, in the order of rent in the set being low). To the top 20%, etc.). Further, the extraction device 100 has a relatively short distance from the nearest station in the product information associated with the property XXX as compared to the set (for example, in the order from the shortest distance from the nearest station in the set to the higher rank). Judgment of superiority, such as belonging to 20%.

  Furthermore, the extraction apparatus 100 acquires the explanatory note F10 uploaded from the business operator as the product information associated with the property XXX. And the extraction apparatus 100 extracts the feature information which is the information which shows the characteristic of a property from the acquired description F10. Specifically, the extraction device 100 inputs the acquired description sentence F10 to the model, and outputs the feature information in the description sentence F10.

  For example, the extraction apparatus 100 uses a model generated corresponding to the category “real estate” of the product, so that “rental”, “corner room”, “from station” or “ "Building age" etc. shall be extracted. Further, the extraction apparatus 100 also extracts the vicinity of the extracted feature information (for example, words included in a predetermined number before and after the explanatory sentence F10 when the morphological analysis is performed) as words related to the feature information. For example, an existing processing method such as important sentence extraction may be applied to such processing.

  Then, the extraction apparatus 100 further specifies a feature in which the property XXX is dominant in the set from the extracted feature information. In the example of FIG. 1, the property XXX is characterized in that items corresponding to “rent” and “time (distance) from the nearest station” are dominant in the set. For this reason, the extraction apparatus 100 determines that these extracted items are feature information in the property XXX, and generates an appeal page that appeals the extracted feature information (step S23).

  Specifically, the extraction device 100 highlights the words and phrases corresponding to the items “rent” and “time (distance) from the nearest station” in the description F10 associated with the property XXX by the business operator. The appeal page of the aspect which did is produced | generated. More specifically, the extraction apparatus 100 generates an appeal page that highlights a word or phrase indicating the extracted feature.

  For example, as shown in FIG. 1, the extraction apparatus 100 sets the location “Fare is a reasonable price of 65,000 yen around here” and the location “Five minutes from the station” in the description F10. A lightly displayed appeal page W10 is generated (in FIG. 1, for the sake of illustration, the relevant part is represented by a frame). Note that the extraction apparatus 100 does not generate a new appeal page W10, but an instruction to highlight the original property introduction page (for example, an instruction that highlights a portion corresponding to feature information) The appeal page W10 may be generated by adding a script described in FIG. In the example of FIG. 1, the appeal page W10 is content in which “rental”, “distance from the station”, and the like are highlighted on the introduction page including the explanatory note F10 and the image P10 of the property XXX.

  Then, the extraction device 100 provides the generated appeal page W10 to the user (step S24). By viewing the appeal page W10, the user can see not only the information described in the description F10 of the property XXX associated with the original business operator but also what excellent features the property XXX has for the set. Can be grasped. Specifically, while viewing the description F10 of the property XXX, the user confirms that the property XXX has relatively low rent and is close to the station among properties satisfying the conditions desired by the user. I can grasp it. That is, when the user browses the page where the property is introduced, the user can efficiently read the feature highlighted by the extracting device 100 without reading all the redundant explanations F10. I can grasp it well.

  As described above, the extraction apparatus 100 according to the embodiment acquires the product information associated with the product that is the target of the commercial transaction on the network. Then, the extraction apparatus 100, for a predetermined product included in the set formed based on the content specified by the user, out of the product information of the predetermined product, information that satisfies a predetermined condition in the set (for example, , Items having specifications belonging to the top 20%, etc.) are extracted. Furthermore, the extraction apparatus 100 generates an appeal page related to a predetermined product based on the extracted feature information.

  Specifically, the extraction apparatus 100 extracts the feature of the property from the product information set by various companies on the property introduction page on the real estate information site, and the appeal page W10 based on the extracted feature. Is generated. As a result, the extraction apparatus 100 generates the appeal page W10 that includes information indicating that the property has particularly excellent characteristics in the set while maintaining information such as the explanatory note F10 provided by the business operator. be able to. As a result, the extraction device 100 can improve the information transferability at the commercial transaction site, and can exhibit the effect that the features relating to the commercial material can be more flexibly appealed.

  Note that the extraction device 100 may use various types of information to complement the learning process. For example, in the learning phase described above, the extraction device 100 uses the appeal page W10 generated by the extraction device 100 to convert the user (for example, making an inquiry about a property or applying for a rental contract). ) Result information (user reaction) may be acquired. Then, the extraction apparatus 100 may feed back to the learning process the feature information given to the appeal page W10 with a good user response such as a high conversion rate as correct data. Thereby, the extraction apparatus 100 can advance learning so as to optimize the process of extracting points that appeal to the user from the explanatory text. Hereinafter, the configuration and the like of the extraction apparatus 100 that performs such processing and the extraction processing system 1 including the extraction apparatus 100 will be described in detail.

[2. Configuration of extraction processing system)
Next, the configuration of the extraction processing system 1 including the extraction device 100 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating a configuration example of the extraction processing system 1 according to the embodiment. As illustrated in FIG. 2, the extraction processing system 1 according to the embodiment includes a user terminal 10, an operator terminal 20, a web server 30, and an extraction device 100. These various devices are communicably connected via a network N by wire or wireless. The extraction processing system 1 illustrated in FIG. 2 may include a plurality of user terminals 10, a plurality of business operator terminals 20, and a plurality of web servers 30.

  The user terminal 10 is, for example, an information processing apparatus such as a smartphone, a desktop PC (Personal Computer), a notebook PC, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or a wearable device. It is. The user terminal 10 accesses the commercial transaction site in accordance with an operation by the user, and performs processing for allowing the user to browse information and apply for a contract for the property at the commercial transaction site.

  The business entity terminal 20 is an information processing terminal used by a business operator who uploads a property on a commercial transaction site. The business entity terminal 20 is, for example, an information processing apparatus such as a smartphone, a desktop PC, a notebook PC, a tablet terminal, a mobile phone, a PDA, or a wearable device.

  The business entity terminal 20 executes various processes for uploading a property to a commercial transaction site in accordance with an operation by the business operator. For example, the business entity terminal 20 uploads text data such as a title of a property and an explanatory text to a web server 30 that provides a commercial transaction site. Further, the business entity terminal 20 uploads the product information including information on the specs (rent etc.) of the property to the web server 30.

  The web server 30 is a server device that provides content (for example, a web page) when accessed from the user terminal 10. In the embodiment, the web server 30 provides a predetermined commercial transaction site (for example, a real estate information site), but in addition, a news site, a weather forecast site, a shopping site, a finance (stock price) site, a route search site, a map providing site Various web pages related to travel sites, restaurant introduction sites, web blogs, etc. may be provided.

  The web page provided by the web server 30 may include an advertisement frame that is a display area for displaying an advertisement and a recommendation frame that is a display area for displaying a recommendation related to a product. Although details will be described later, the extraction device 100 may generate feature information of a product displayed in an advertisement frame or a recommendation frame.

  The extraction device 100 acquires the product information associated with the product, and in the predetermined product included in the set formed based on the content specified by the user, the product information of the predetermined product Among them, feature information that is information satisfying a predetermined condition in the set is extracted. Furthermore, the extraction apparatus 100 generates an appeal page related to a predetermined product based on the extracted feature information.

  And the extraction apparatus 100 provides the produced | generated appeal page to the user terminal 10, when access is received from a user. In addition, the extraction apparatus 100 according to the embodiment may also serve as a configuration as the web server 30 that manages and manages a commercial transaction site or manages information uploaded from a business operator. That is, the extraction device 100 and the web server 30 may be separate devices or may be realized by a device having both functions.

[3. (Extractor configuration)
Next, the configuration of the extraction apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating a configuration example of the extraction device 100 according to the embodiment. As illustrated in FIG. 3, the extraction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The extraction device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator who uses the extraction device 100, and a display unit (for example, a liquid crystal display) for displaying various types of information. You may have.

(About the communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to the network N in a wired or wireless manner, and transmits / receives information to / from the user terminal 10, the provider terminal 20, and the web server 30 via the network N.

(About the storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 includes a product information storage unit 121, a learning information storage unit 122, and a provision information storage unit 123.

(About the product information storage unit 121)
The merchandise information storage unit 121 stores information related to merchandise uploaded to the commercial transaction site. Here, FIG. 4 shows an example of the product information storage unit 121 according to the embodiment. FIG. 4 is a diagram illustrating an example of the product information storage unit 121 according to the embodiment. In the example illustrated in FIG. 4, the product information storage unit 121 includes items such as “product ID” and “product information”. In addition, “Product Information” includes “Category”, “Region”, “Floor Layout”, “Rent”, “Time from Nearest Station”, “Area”, “Age”, “Pet”, “Explanation” ”.

  “Product ID” indicates identification information for identifying the product. In the embodiment, the identification information such as the product ID is common to the reference numerals used in the description. For example, a product whose product ID is “B01” may be expressed as “product B01”.

  “Product information” indicates information associated with the product. For example, the product information includes an item indicating the specification of the product, a value for each item, and the like. In addition, the product information includes information such as an explanatory note set by a company that is a provider of the product and a title of the product.

  The product information includes different information for each product category. For example, when the product category is “real estate”, items such as “rent” and “area” are included in the product information. Alternatively, when the category of the product is “PC”, items such as “CPU” and “memory capacity” are included in the product information.

  “Category” indicates a category to which a commercial item on the commercial transaction site belongs. “Region” indicates an area where the product (a property in the example of FIG. 4) is located. “Floor layout” indicates the floor plan of the property. “Rent” indicates the rent of the property. “Time from the nearest station” indicates the distance from the nearest station or the travel time required from the nearest station. “Area” indicates the area occupied by the property. “Building age” indicates the number of years that have elapsed since the property was built. “Pet” indicates information such as whether or not there is a property that allows pet breeding.

  The “description” indicates an explanation associated with the property. In the example shown in FIG. 4, “description” is indicated by a concept such as “F01”, but actually, text data constituting the description is stored in the item of the description.

  That is, in the example of the data shown in FIG. 4, the category of the product in the product B01 identified by the product ID “B01” is “real estate”, and the product information of the product B01 includes the region where the product is located. It is “AAA Ward”, the floor plan is “1K”, the rent is “65000” yen, the time from the nearest station is “5 minutes on foot”, the area is “22” square meters, and pet breeding Is “OK”, and the explanatory text is “F01”.

(About the learning information storage unit 122)
The learning information storage unit 122 stores information related to learning processing based on the product information. Here, FIG. 5 illustrates an example of the learning information storage unit 122 according to the embodiment. FIG. 5 is a diagram illustrating an example of the learning information storage unit 122 according to the embodiment. In the example illustrated in FIG. 5, the learning information storage unit 122 includes items such as “learning data ID”, “learning data”, “category”, and “extraction item”.

  “Learning data ID” indicates identification information for identifying learning data. “Learning data” indicates the content of learning data actually used in the learning process. For example, the learning data according to the embodiment is product information associated with a product.

  “Category” indicates a category to which the product belongs. For example, the extraction device 100 performs learning for each category of merchandise and generates a model for each category. “Extracted item” indicates an item learned to be extracted as feature information in each category.

  That is, in the example of the data shown in FIG. 5, the learning data C01 identified by the learning data ID “C01” is learned using the product “B01”, the product “B02”, the product “B03”, and the like as learning data. The category is “real estate”, and the extracted items as feature information are, for example, “rent”, “area”, “time from the nearest station”, “building age” It is shown that.

Note that the learning information storage unit 122 may store a model learned based on the learning data. Further, the extraction device 100 may update the learning data stored in the learning information storage unit 122 as appropriate and continue learning the generated model.

  In the example shown in FIG. 5, “rent” or the like is shown as the extraction item. However, the extraction apparatus 100 is assumed not to extract only the term “rent” but to indicate “rent”. Synonyms and the like may be stored as extracted items. For example, the extraction device 100 estimates that terms such as “rent” and “monthly payment” are synonymous by manual setting or learning processing. In this case, the extraction device 100 does not include a clear phrase “rent” in the explanatory text, but when the phrase “monthly payment is ...” appears, the phrase is “rent”. Can be extracted as feature information corresponding to.

  Although not shown in FIG. 5, items of learning data include not only product information used for learning, but also information manually set as items to be extracted as feature information or explanatory text, for example. Correct answer data such as a summary sentence manually created based on this may be included. Further, in the case of using a deep learning method in learning, it is not always necessary to use only learning data for which a correct example (correct answer data) is clearly identified. For example, the extraction device 100 does not necessarily need correct data as long as it can learn the feature of text data (document) included in the product information and recognize the feature information included in the text data.

(Provided information storage unit 123)
The provided information storage unit 123 stores information related to a page provided to the user and generated by the extraction device 100 (appeal page). Here, FIG. 6 illustrates an example of the provision information storage unit 123 according to the embodiment. FIG. 6 is a diagram illustrating an example of the provision information storage unit 123 according to the embodiment. In the example illustrated in FIG. 6, the provided information storage unit 123 includes “appeal page ID”, “original product ID”, “collection information”, “extraction item”, “appeal point”, “effect measurement information”, It has items such as “review information”.

  “Appeal page ID” indicates identification information for identifying an appeal page generated by the extraction apparatus 100. The “original merchandise ID” indicates identification information of the merchandise that is the origin of the appeal page.

  “Set information” indicates information related to the set formed by the user. In the example shown in FIG. 6, “collection information” is indicated by a concept such as “F01”, but actually, various specific information is stored. For example, the set information includes conditions specified by the user, identification information of a plurality of products included in the set formed based on the conditions specified by the user, and the like. Note that the set information is dynamic information. In other words, the set changes every time a condition designated by the user is changed, so that the set information also changes every time the condition designated by the user is changed. In addition, as the set changes, the superior information in the set also changes, so the items (feature information) highlighted on the appeal page also change.

  The “extraction item” corresponds to the same item shown in FIG. “Appeal point” indicates an item that is dominant in the set among items extracted as information characterizing the product. For example, when the price of rent is superior to other products among a plurality of products included in the set, the appeal point includes the item “rent”. As described above, the condition such as which item is “dominant” may be artificially set by an administrator of the extraction apparatus 100, for example.

  “Effect measurement information” indicates result information indicating that the appeal page exerts an effect among the information on the user's reaction to the appeal page. “Review information” indicates information related to a review posted from the user to the appeal page among information related to the user's reaction to the appeal page. In the example shown in FIG. 6, “effect measurement information” and “review information” are represented by concepts such as “G01” and “H01”, but actually, various specific information is stored. The

  For example, the “effect measurement information” may store the number of times that the appeal page has been accessed by the user and the number and rate of users who have reached conversion on the appeal page. Examples of conversion include an application for a preview of the property on the appeal page, and an access to a property introduction page (for example, a web page of a company that provides the property). In addition, the effect measurement information may store, for example, the number of items selected by the user (the number of clicks or touches) or the rate without reaching a specific action such as a preview. Further, the effect measurement information may store the frequency with which the user accesses the appeal page, the time spent on the appeal page, and the like. That is, any information may be stored in the effect measurement information as long as it is information related to the user's response to the appeal page.

  Further, “review information” stores information related to user reviews for appeal pages. For example, in the appeal page, when a system in which a user evaluates a property with a numerical value in five stages is adopted, the score evaluated by the user is stored in the review information. For example, the review information stores an average value of numerical values from “0” to “5” transmitted from the user. The review information may store the number and rate of user reviews transmitted from the user, or specific user review comments (text data). That is, any information may be stored in the review information as long as it is information related to the user review for the appeal page.

  That is, the example of the data shown in FIG. 6 indicates that the appeal page W01 identified by the appeal page ID “W01” is a page related to the product B01 identified by the original product ID “B01”. Yes. Also, the set information associated with the appeal page W01 is “F01”, and the extracted items are “rent”, “area”, “time from the nearest station”, “building age”, etc. Of these, items that are the appealing points that are superior to the set are “rent” and “time from the nearest station”. Further, the effect measurement information of the appeal page W01 is “G01”, and the review information is “H01”.

(About the control unit 130)
The control unit 130 is, for example, a controller, and includes various programs (an example of an extraction program) stored in a storage device inside the extraction device 100 by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. This is realized by executing the RAM as a work area. The control unit 130 is, for example, a controller, and is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As shown in FIG. 3, the control unit 130 includes an acquisition unit 131, a learning unit 132, a reception unit 133, an extraction unit 134, a generation unit 135, and a provision unit 136. Information described below Implement or execute a processing function or action. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 3, and may be another configuration as long as the information processing described below is performed. Further, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 3, and may be another connection relationship.

(About the acquisition unit 131)
The acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires the product information associated with the product that is the target of the commercial transaction on the network.

  For example, the acquisition unit 131 acquires an item indicating the spec of the product and a value (numerical value or content) corresponding to the item. The acquisition unit 131 acquires different product information for each product category. As an example, if the category of the product is real estate, the acquisition unit 131 acquires the product information shown in FIG. 4 such as the rent and the area as the product information. The acquisition unit 131 may acquire any information as product material information as long as it is information such as specifications that characterize the product.

  Then, the acquisition unit 131 associates the product with the product information and stores it in the product information storage unit 121 as illustrated in FIG. That is, the acquisition unit 131 constructs a database related to merchandise at the commercial transaction site. In such a database, for example, it is possible to perform data management that sorts properties according to rent, size, etc., or extracts the top 20% when sorted.

  Further, the acquisition unit 131 acquires text data associated with the product as the product information. For example, the acquisition unit 131 acquires text data corresponding to an explanatory note for explaining a product on a page introducing the product. In addition, the page which introduces merchandise includes a page where an application for merchandise can be performed, and a page where transactions such as purchase and sale of merchandise can be performed.

  For example, when the product is a property, the acquisition unit 131 performs a morphological analysis on the explanatory text associated with the property, and acquires a word included in the explanatory text. Furthermore, the acquisition part 131 acquires the word which shows the characteristic of a property. More specifically, the acquisition unit 131 acquires, for example, a word indicating a rent, an area, an age, a distance, a nearest station, or the like as a word indicating the feature of the property. Note that the acquisition unit 131 is not limited to the explanatory text, and may acquire any information as long as it is text data associated with a commercial material such as a title.

  In addition, when the morphological analysis is performed on the title or the explanatory text, the acquiring unit 131 may generate an appearance frequency of words or a plurality of words based on the number of appearances of words (phrases) included in the document (here, the title or the explanatory text). You may calculate the importance of each word in the description sentence. For example, the acquisition unit 131 may calculate a DF value of a word related to the acquired explanatory text. Moreover, the acquisition part 131 may acquire not only DF value but index values, such as tf-idf. Then, the acquisition unit 131 may acquire only words that exceed a predetermined threshold. Thereby, the acquisition part 131 can acquire only the word assumed to characterize a property, without acquiring the common phrase etc. which are contained in various documents.

  Moreover, the acquisition part 131 may acquire the information regarding the reaction with respect to the said appeal page from the user who actually browsed the appeal page. For example, the acquisition unit 131 acquires at least one of the number or rate at which the product is selected from the user or the number or rate of conversions related to the product as the user's response to the appeal page or the product. Also good. Moreover, the acquisition part 131 may acquire the information regarding the review transmitted from a user. Then, the learning unit 132, the extraction unit 134, and the generation unit 135, which will be described later, may appropriately tune the generated model and the feature information to be extracted based on the information about the user response acquired by the acquisition unit 131. Good.

  Note that the acquisition unit 131 may acquire information regarding a user who has viewed the appeal page generated by the generation unit 135. The information about the user is, for example, user attribute information (gender, age, place of residence, occupation, etc.) or user behavior information (past browsing history, purchase history, etc. on the network). For example, the acquisition unit 131 specifies a user based on a cookie transmitted from the user terminal 10 and acquires information regarding the specified user.

  The acquisition unit 131 stores the acquired information in each unit of the storage unit 120 as appropriate. The acquisition unit 131 may appropriately acquire information used by each processing unit described later from each unit of the storage unit 120.

(About the learning unit 132)
The learning unit 132 performs learning for extracting feature information that is information indicating the feature of the product from the product information acquired by the acquisition unit 131.

  More specifically, the learning unit 132 performs learning for extracting feature information indicating features of the product from text data (description) set in advance by the business with respect to the product. .

  For example, the learning unit 132 performs learning for extracting feature information based on learning processing (for example, supervised machine learning) using correct answer data set in advance as items to be extracted from the product information. For example, the learning unit 132 acquires, as correct data, information obtained by artificially extracting items and phrases that are assumed to indicate the characteristics of the product. And the learning part 132 learns the feature information which should be extracted from merchandise information by learning these correct answer data (sample).

  Note that the learning unit 132 may perform learning for extracting feature information based on a predetermined condition set in advance for each product category. Specifically, when the product is “real estate”, the learning unit 132 receives items such as “rent” and “area” as correct data as feature information to be extracted. Then, the learning unit 132 learns which part of the explanatory text corresponds to an item such as “rent” or “area” based on the correct answer data. In this way, the learning unit 132 can efficiently learn which information in the explanatory text indicates the characteristics of the product by using the correct answer data for each category as a sample.

  In addition, the learning unit 132 may extract feature information using a model (learning device) that has been learned to extract information characterizing the explanatory text associated with the product from the product information. Good. For example, by using a technique such as deep learning, the learning unit 132 can generate a model for extracting information (feature amount) that characterizes each explanatory sentence without having clear correct answer data.

  The learning unit 132 may generate a model using any of the learning processes described above. Further, the learning unit 132 may appropriately use known learning processes other than those exemplified above. That is, the learning method is limited if the learning unit 132 can perform feature information extraction processing by the extraction unit 134 described later and appeal page generation processing by the generation unit 135. Absent.

  Note that the learning unit 132 may perform a process of adjusting feature information to be extracted based on information related to a user's reaction. For example, the learning unit 132 has extracted features in an appeal page (for example, a page in which conversion exceeds a predetermined number or a page in which a user review score exceeds a predetermined score) in which a response from the user exceeds a predetermined threshold. Information is determined as correct data, and the model is updated. That is, the learning unit 132 treats the appeal page with relatively good response from the user as correct data in model generation, assuming that the feature information can be extracted appropriately. Thereby, the learning unit 132 can optimize the model.

  The learning unit 132 may generate a model that extracts different feature information for each user. For example, in the feature information appealing on the appeal page, there may be feature information whose appeal effect varies depending on the user's attributes. For example, in an appeal page in which a certain feature is highlighted, there is a possibility that the conversion rate is high for female users while the conversion rate is low for male users. In such a case, the learning unit 132 may perform learning so as to adjust items extracted from the merchandise information according to the user attributes. Specifically, the learning unit 132 may perform learning to adjust items extracted as feature information so that the appeal page for female users and the appeal page for men are displayed in different modes. Good. As a result, the generation unit 135 described later generates different appeal pages that emphasize each feature information that is assumed to have a high appeal effect to each user even if the pages appeal the same product. can do.

  The learning unit 132 appropriately stores the learning data used for learning and the model corresponding to the learning data in the learning information storage unit 122.

(About the reception unit 133)
The accepting unit 133 accepts various requests related to commercial transactions. For example, the reception unit 133 receives a request from the user terminal 10 to search for a commercial material that meets the conditions desired by the user at the commercial transaction site.

  Specifically, the reception unit 133 receives a search query for searching for a property from the user terminal 10 that has accessed the commercial transaction site. In addition, the reception unit 133 receives a browsing (acquisition) request for a page related to a product desired by the user from a user who has browsed a search result page corresponding to the search query. That is, the reception unit 133 receives an access request for an appeal page desired by the user. The accepting unit 133 may accept various requests directly from the user terminal 10 via the web server 30.

(About the extraction unit 134)
The extraction unit 134 is information on a predetermined product included in a set formed based on contents specified by the user, and satisfies a predetermined condition in the set among the product information of the predetermined product. Extract feature information.

  Specifically, the extraction unit 134 extracts the product information as feature information when the product information of a predetermined product included in the set is information indicating superiority in the set as a predetermined condition. To do.

  For example, the extraction unit 134 extracts feature information based on a predetermined condition set in advance for each category of merchandise. Specifically, the extraction unit 134 first extracts characteristic items from the product information using the model generated by the learning unit 132. Then, the extraction unit 134 extracts, as feature information, items that satisfy a predetermined condition set in advance for each category from among the extracted items as information indicating superiority in the set.

  As an example, if the category is real estate, the extraction unit 134 may be “items of product information of a plurality of properties that meet the conditions searched by the user (that is, a set formed based on conditions specified by the user). Among them, a condition such as “information belonging to the top 20% is extracted as feature information” is received in advance. And the extraction part 134 extracts the item which adapts the said conditions among the extracted items as feature information. In the example of FIG. 1, the extraction unit 134 determines that two items, “rent” and “time to the nearest station”, among the product information of the property XXX that the user tried to browse, have met the above conditions. judge. Then, the extraction unit 134 extracts these items as feature information (appeal points).

  Note that information that has superiority in a set does not necessarily have to be ranked information. For example, the information having superiority in the set may be information indicating an advantage that the other product does not necessarily have in the set. Specifically, when the product is a property, it can be said that the information that “pet breeding is possible” is information indicating an advantage that other products do not necessarily have. In this case, as the predetermined condition, for example, “a predetermined specification that does not have more than half of the items in the item information items of a plurality of properties that match the condition searched by the user is extracted as feature information. If the setting is “Yes”, the extraction unit 134 may extract information indicating that “pet breeding is possible” as feature information.

  Note that the extracting unit 134 determines conformity with the above conditions by referring to a database constructed from the product information acquired by the acquiring unit 131.

  Here, the extraction unit 134 extracts feature information from the description text or the like (text data) of the product among the product information associated with the product. Such processing is realized by using the model generated by the learning unit 132 described above.

  That is, the extraction unit 134 extracts feature information based on learning processing using correct data set for text data associated with a product among the product information. In this case, the extraction unit 134 inputs text data to a model that has been artificially given correct data in advance and learned, so that any part of the text data corresponds to the “item to be extracted”. Can be determined. And the extraction part 134 extracts the item which satisfy | fills the above predetermined conditions among the items extracted from text data as the feature information of the said commercial material.

  Further, the extraction unit 134 may extract feature information using a model (learning device) learned to extract information characterizing text data associated with the product from the product information. Such processing is realized by using a model learned using a technique such as deep learning.

  That is, the extraction unit 134 extracts words and sentences determined to characterize the text data associated with the product from the product information. Then, the extracting unit 134 associates the extracted word or sentence with each item included in the product database, so that any part of the text data corresponds to the “item to be extracted”. Determine whether. Note that such processing for associating the extracted phrases and sentences with each item included in the product database is also realized by using a model learned using the above-described deep learning method. The And the extraction part 134 extracts the item which satisfy | fills the above predetermined conditions among the items extracted from text data as the feature information of the said commercial material.

  For example, the extraction unit 134 extracts a phrase that characterizes the product as the feature information. In other words, the extraction unit 134 extracts words that characterize the merchandise and that show superiority in the set. The generation unit 135 described later generates an appeal page by highlighting a word (appeal point) extracted by the extraction unit 134.

  Note that the feature information (words) extracted by the extraction unit 134 is not limited to one word such as “rent” or “area”. For example, the extraction unit 134 may extract a sentence that is assumed to describe “rent” as the feature information. Further, the extraction unit 134 may extract a word / phrase that is assumed to describe “rent” and a predetermined number of words before and after it as feature information. Such processing may be realized by using known important sentence extraction processing or the like. As a result, the extraction unit 134 does not become unnatural for the user even when extracting a phrase that is supposed to describe “rent” and a predetermined number of words before and after that. Can be selected and extracted.

  Note that the extraction unit 134 may perform processing for adjusting feature information to be extracted based on information related to a user's reaction. For example, the extraction unit 134 may extract features extracted from appealing pages (for example, pages where conversions exceed a predetermined number or pages where user reviews have exceeded a predetermined number) in which the reaction from the user exceeds a predetermined threshold. The information is determined as correct data, and the result is fed back to the learning unit 132. In other words, the extraction unit 134 treats the appeal page with relatively good response from the user as the correct answer data in the model generation, assuming that the feature information is appropriately extracted. Thereby, the extraction unit 134 can optimize the extraction process.

  Further, the extraction unit 134 may extract different feature information for each user. As described in the learning unit 132, the appeal point in the product may change the appeal effect according to the attribute of the user. For this reason, the extraction unit 134 may adjust the feature information to be extracted for the product based on, for example, the correlation between the user attribute and the user reaction.

(About the generator 135)
The generation unit 135 generates a page related to a predetermined product based on the feature information extracted by the extraction unit 134.

  For example, the generation unit 135 generates a page in which the word / phrase (feature information) extracted by the extraction unit 134 is highlighted in the text data associated with the product. Specifically, the generation unit 135 generates an appeal page in which the portion corresponding to the feature information in the text data is highlighted and displayed. Thereby, the production | generation part 135 can appeal to a user immediately the information which shows the specification about how much the characteristic displayed with respect to a group | goods displayed on the user terminal 10 has.

  The generation unit 135 may generate an appeal page based on a reaction from the user. That is, similar to the extraction unit 134 described above, the generation unit 135 also learns the generation process based on the user's reaction, using the highlighted range, the display mode of the feature information, etc. as correct data (or incorrect data). The generated appeal page may be optimized.

  As a specific example, a case where a property is converted will be described. In the appeal page, if the highlighted part of the information presented as feature information (for example, the text corresponding to the part explaining the rent) is not appropriate, or the method of highlighting is not appropriate, in other words, When the appeal point of the product is not well communicated to the user, the user is assumed to ask for the application for the property. On the other hand, when the property is converted from the user, it is assumed that the feature information and the display method appealed on the appeal page are correct.

  Therefore, the generation unit 135 estimates the appeal page mode and the correctness of the feature information based on the user's reaction. For example, the generation unit 135 sets a predetermined threshold value in advance. For example, the generation unit 135 sets a threshold for the conversion rate related to a certain property. And the production | generation part 135 determines whether the conversion rate exceeding a threshold value was achieved within the predetermined period or during a fixed number of browsing.

  When the conversion rate exceeds the threshold, the generation unit 135 assumes that the appeal page related to the property is correct and feeds back the appeal page as one of the positive examples. Further, when the conversion rate does not exceed the threshold, the generation unit 135 assumes that the appeal page related to the property is incorrect and feeds back the appeal page as one of negative examples. Thus, by continuing learning using the user's reaction, the generation unit 135 causes the feature information to be appealed, page coloration, and highlight processing (for example, font size, font size, Type, layout of the entire page, etc.).

  In addition, as a method of obtaining such a user's reaction, the extraction apparatus 100 may perform manual determination using, for example, crowdsourcing. By performing the learning using the manpower, the generation unit 135 can generate an appealing page that is not unnatural when viewed by a person.

  Moreover, the production | generation part 135 may produce | generate a different appeal page for every user based on the information regarding a user. That is, like the extraction unit 134 described above, the generation unit 135 may generate an appeal page that is assumed to be more effective for the user based on the user's attribute information, action history, and the like.

(About the providing unit 136)
The providing unit 136 provides various information. For example, when the receiving unit 133 receives a merchandise browsing request, the providing unit 136 provides (transmits) an appeal page to the user who transmitted the request. Further, when the receiving unit 133 receives a search query for searching for a product, the providing unit 136 provides the user terminal 10 with a search result corresponding to the received search query.

[4. Web server configuration]
Next, the configuration of the web server 30 according to the embodiment will be described with reference to FIG. FIG. 7 is a diagram illustrating a configuration example of the web server 30 according to the embodiment. As illustrated in FIG. 7, the web server 30 includes a communication unit 31, a content storage unit 32, and a control unit 33.

  The communication part 31 is implement | achieved by NIC etc., for example. The communication unit 31 is connected to the network N in a wired or wireless manner, and transmits / receives information to / from the user terminal 10, the provider terminal 20, and the extraction device 100.

  The content storage unit 32 is realized by, for example, a semiconductor memory device such as a RAM or a flash memory, or a storage device such as a hard disk or an optical disk. And the content memory | storage part 32 memorize | stores the web page which is an example of a content. For example, the content storage unit 32 stores an HTML file that forms a web page, a still image or a moving image displayed on the web page. The web page stored in the content storage unit 32 may include an advertisement acquisition command for acquiring advertisement content to be displayed on the web page.

  The control unit 33 is a controller, and is realized, for example, by executing various programs stored in a storage device inside the web server 30 using the RAM as a work area by a CPU, an MPU, or the like. The control unit 33 is a controller and is realized by an integrated circuit such as an ASIC or FPGA, for example.

  As illustrated in FIG. 7, the control unit 33 includes a reception unit 34 and a distribution unit 35, and realizes or executes information processing functions and operations described below. The internal configuration of the control unit 33 is not limited to the configuration illustrated in FIG. 7, and may be another configuration as long as the information processing described later is performed. In addition, the connection relationship between the processing units included in the control unit 33 is not limited to the connection relationship illustrated in FIG. 7, and may be another connection relationship.

  The accepting unit 34 accepts a web page acquisition request from the user terminal 10. For example, the reception unit 34 receives an HTTP request as a web page acquisition request. In addition, the reception unit 34 receives information related to the product from the business entity terminal 20. The receiving unit 34 stores information related to the merchandise received from the business entity terminal 20 in the content storage unit 32.

  The distribution unit 35 distributes the web page to the user terminal 10 when the reception unit 34 receives a web page acquisition request. Specifically, the distribution unit 35 acquires an acquisition request target web page from the content storage unit 32 and distributes the acquired web page to the user terminal 10. As described above, the web page stored in the content storage unit 32 includes an advertisement acquisition command. That is, when displaying the acquired web page, the user terminal 10 transmits an advertisement content distribution request to the extraction device 100 in accordance with the advertisement acquisition command included in the web page.

  In the above-described embodiment, an example in which an appeal page is provided by the providing unit 136 related to the extraction device 100 has been shown. However, a web page such as an appeal page is distributed to the user terminal 10 by the distribution unit 35 related to the web server 30. May be. In this case, the web server 30 cooperates with the extraction apparatus 100 to perform processing for transmitting information necessary for generating an appeal page to the extraction apparatus 100 and acquiring an appeal page from the extraction apparatus 100. .

[5. Processing procedure)
Next, a processing procedure performed by the extraction apparatus 100 according to the embodiment will be described with reference to FIGS. 8, 9, and 10. First, a processing procedure related to model generation will be described with reference to FIG. FIG. 8 is a flowchart (1) illustrating a processing procedure according to the embodiment.

  As illustrated in FIG. 8, the extraction device 100 acquires product information associated with the extraction device product as learning data for generating a model (step S101).

  Then, the extraction apparatus 100 generates a database based on the acquired product information (step S102). Further, the extraction device 100 generates a learning model based on the acquired product information (step S103). Extraction Device According to the above flow, the learning phase by the extraction device 100 ends.

  Next, a processing procedure related to extraction processing using a model will be described with reference to FIG. FIG. 9 is a flowchart (2) illustrating a processing procedure according to the embodiment.

  As illustrated in FIG. 9, the extraction device 100 determines whether or not designation regarding a product has been received from the user terminal 10 (step S <b> 201). If the extraction device 100 has not received a designation regarding a product (step S201; No), the extraction device 100 stands by until it is received.

  On the other hand, when designation regarding a product is received (step S201; Yes), the extraction apparatus 100 identifies a set corresponding to the received designation (condition) (step S202). Then, the extraction apparatus 100 extracts feature information having an advantage over the set in a predetermined product included in the set (step S203).

  Subsequently, the extraction apparatus 100 generates an appeal page based on the extracted feature information (step S204). Then, the extraction apparatus 100 provides an appeal page to the user who has specified the product (Step S205). With the above flow, the generation phase by the extraction device 100 ends.

  Next, the feedback procedure regarding the extraction process and the generation process will be described with reference to FIG. FIG. 10 is a flowchart (3) illustrating the processing procedure according to the embodiment.

  As illustrated in FIG. 10, the extraction device 100 determines whether a response to the appeal page has been received from the user terminal 10 (step S301). If the extraction device 100 has not received a reaction (step S301; No), the extraction device 100 stands by until it is received.

  On the other hand, when a reaction is received (step S301; Yes), the extraction apparatus 100 acquires a reaction from the user (step S302). Then, the extraction device 100 determines whether or not the reaction from the user has exceeded a predetermined threshold (step S303).

  When the reaction from the user exceeds a predetermined threshold (step S303; Yes), the extraction apparatus 100 determines that the feature information (or the appeal page itself) on the appeal page provided to the user is correct (step). S304). On the other hand, when the reaction from the user does not exceed the predetermined threshold (step S303; No), the extraction device 100 determines that the feature information is incorrect (step S305).

  Then, the extraction apparatus 100 updates the model using learning data in which the data obtained in steps S304 and S305 is a positive example or a negative example (step S306). In other words, the extraction device 100 learns the model more strongly based on the user's reaction. By repeating the above flow, the extraction device 100 optimizes the model so that feature information can be extracted more appropriately.

[6. (Modification)
The extraction device 100 described above may be implemented in various different forms other than the above embodiment. Therefore, in the following, another embodiment of the extraction device 100 will be described.

[6-1. (Type of feature information)
In the said embodiment, the extraction apparatus 100 showed the example which produces | generates an appeal page in the aspect which highlights the location corresponding to characteristic information among the explanatory notes of commerce. Here, the extraction apparatus 100 may generate an appeal page having a mode different from a mode in which the feature information is highlighted.

  For example, the extraction device 100 may generate a summary sentence of the explanatory text associated with the product based on the feature information, and may generate an appeal page to which the summary text is given. Specifically, the extraction apparatus 100 generates an appeal page in a form in which the generated summary text is arranged prior to the explanatory text previously associated with the product by the business operator. That is, the extraction apparatus 100 can transmit a summary sentence that simply conveys only the characteristics of the product such as a point superior to the set to the user before the explanatory sentence. Thereby, even if the explanatory text is redundant, for example, the user can accurately grasp the characteristics of the product without reading the explanatory text itself.

[6-2. (Use for search results)
The extraction apparatus 100 may not only generate an appeal page using the generated feature information but also use the feature information for various processes. For example, the extraction apparatus 100 may display the feature information together with the search result in the search for the commercial material performed at the commercial transaction site or the like.

  In other words, the extraction device 100 may generate a search result page including product feature information when a product is displayed as a result of a predetermined search process. In this case, the search result page displays, for example, characteristic information (for example, summary text) related to the product in the vicinity of the product searched based on the search query. For this reason, the user can grasp the characteristics of the product without displaying the appeal page when searching for the product. Thus, the extraction device 100 can improve the convenience of the user who uses the commercial transaction site.

[6-3. (Use for recommendation)
The extraction apparatus 100 displays the feature information not only in the commercial transaction site but also in, for example, a recommendation of a product displayed together with a web page or the like (for example, information displayed in an advertisement frame or a recommendation frame of a web page). May be.

  For example, there is a case where a product that is exhibited on a commercial transaction site is displayed as a recommendation on a web page or the like other than the product. In this case, the extraction apparatus 100 displays the feature information together with the recommended product. Accordingly, the user can grasp the characteristics of the product without accessing the appeal page of the recommended product (without changing the display from the currently viewed page to the appeal page).

  Note that the extraction apparatus 100 may acquire information indicating whether a recommendation for the product has been clicked or whether the user has converted. For example, when the feature information displayed as a recommendation is inappropriate or the appeal effect is low, it is assumed that the user does not return a click or conversion response. On the other hand, when the user returns a response that the user clicks or converts, the feature information displayed together with the recommendation is assumed to be information that appropriately indicates the feature of the product (correct answer).

  Then, as described in the above embodiment, the extraction apparatus 100 determines whether or not the feature information or the feature information processing is correct according to the reaction to the recommendation. Thereby, the extraction apparatus 100 can improve the accuracy of the extraction process or the extraction process (learning) using a reaction from a user other than the appeal page.

[6-4. commodity〕
In the above-described embodiment, a rental property of a real estate, a personal computer, or the like is given as an example of the product, but the product is not limited to these. That is, the extraction device 100 performs the extraction process according to the embodiment for any product such as a product or a service as long as it is a product handled at a commercial transaction site and can acquire product information. be able to.

[6-5. content〕
In the said embodiment, although the web page was mentioned as an example as content, a content is not restricted to this example. For example, the content may be advertising content, an information page provided via an application, or content related to a recommendation. That is, the content generated by the extraction device 100 may be any content as long as it is information content provided via a network.

  For example, when extracting the feature information of a predetermined product, the extraction apparatus 100 may reflect the extracted feature information on advertising content for promoting the product. For example, the extraction device 100 extracts feature information about a product advertised by an advertising content submitted by an advertiser (product provider). Then, the extraction device 100 adds advertisement information (for example, a summary sentence) generated from the extracted feature information to the advertisement content to generate the advertisement content. Specifically, the extraction device 100 adds a summary sentence generated based on the feature information to a place where text data for promoting a product is inserted in the advertisement content. Thereby, the extraction apparatus 100 can generate, for example, advertisement content that appeals more advantageous features to many competing products (that is, a set of products).

[7. Hardware configuration)
The extraction device 100, the user terminal 10, the provider terminal 20, and the web server 30 according to the embodiments described above are realized by a computer 1000 having a configuration as shown in FIG. 11, for example. Hereinafter, the extraction apparatus 100 will be described as an example. FIG. 11 is a hardware configuration diagram illustrating an example of a computer 1000 that implements the functions of the extraction device 100. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores a program executed by the CPU 1100, data used by the program, and the like. The communication interface 1500 receives data from other devices via the communication network 500 (corresponding to the network N shown in FIG. 2) and sends the data to the CPU 1100. Send to device.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. Further, the CPU 1100 outputs the data generated via the input / output interface 1600 to the output device.

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  For example, when the computer 1000 functions as the extraction device 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing a program loaded on the RAM 1200. The HDD 1400 stores data in the storage unit 120. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from other devices via the communication network 500.

[8. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified. For example, the various types of information illustrated in each drawing is not limited to the illustrated information.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. For example, the acquisition unit 131 and the reception unit 133 illustrated in FIG. 3 may be integrated. Further, for example, information stored in the storage unit 120 may be stored in a predetermined storage device provided outside via the network N.

  Moreover, in the said embodiment, the extraction apparatus 100 showed the example which performs the acquisition process which acquires merchandise information, the extraction process which extracts feature information, and the production | generation process which produces | generates an appeal page, for example. However, the extraction apparatus 100 described above may be separated into an acquisition apparatus that performs an acquisition process, an extraction apparatus that performs an extraction process, and a generation apparatus that performs a generation process. In this case, the acquisition device includes at least the acquisition unit 131. The extraction device has at least an extraction unit 134. The generation device has at least a generation unit 135. And the process by said extraction apparatus 100 is implement | achieved by the extraction processing system 1 which has each apparatus of an acquisition apparatus, an extraction apparatus, and a production | generation apparatus.

  In addition, the above-described embodiments and modifications can be combined as appropriate within a range that does not contradict processing contents.

[9. effect〕
As described above, the extraction device 100 according to the embodiment includes the acquisition unit 131 and the extraction unit 134. The acquisition unit 131 acquires the product information associated with the product that is the target of the commercial transaction on the network. The extraction unit 134 is characteristic information that is information satisfying a predetermined condition in the set, out of the product information of the predetermined product, in the predetermined product included in the set formed based on the content specified by the user. To extract.

  As described above, the extraction apparatus 100 according to the embodiment extracts the feature of the product from the product information set by various business operators, for example, on the product introduction page on the commercial transaction site. That is, the extraction apparatus 100 appeals dynamic features according to the user's specification by extracting the appeal points that are characteristic when the product is compared with the set (other products) specified by the user. In other words, it is possible to appeal the features relating to the commercial product more flexibly.

  Further, the extraction unit 134 extracts the product information as feature information when the product information of a predetermined product included in the set is information indicating superiority in the set.

  As described above, the extraction apparatus 100 according to the embodiment extracts features having superiority when compared with the set designated by the user, and thus can clearly convey the features of the commercial material to the user.

  Further, the extraction unit 134 extracts feature information based on a predetermined condition set in advance for each category of merchandise.

  Thus, the extraction apparatus 100 according to the embodiment sets conditions for each category and extracts feature information. Thereby, the extraction apparatus 100 can extract accurate feature information corresponding to various products.

  Further, the extraction unit 134 extracts feature information based on learning processing using correct data set for text data associated with a product among the product information.

  As described above, the extraction apparatus 100 according to the embodiment can extract feature information similar to a feature extracted manually by performing learning based on preset correct answer data. That is, the extraction apparatus 100 can accurately extract feature information indicating the feature of the product.

  Further, the extraction unit 134 extracts feature information using a learning device that has been learned to extract information that characterizes text data associated with the product from the product information.

  As described above, the extraction apparatus 100 according to the embodiment may perform learning by a method of extracting feature quantities characterizing text data by a method such as deep learning. By such processing, the extraction apparatus 100 can extract feature information with high accuracy and efficiency.

  In addition, the extraction apparatus 100 according to the embodiment further includes a generation unit 135 that generates a page (an example of content) related to a predetermined product based on the feature information extracted by the extraction unit 134.

  As described above, the extraction apparatus 100 according to the embodiment extracts features of a product, for example, on a product introduction page on a commercial transaction site, and generates a page for appealing the feature based on the extracted feature. Thereby, the extraction apparatus 100 can appeal the dynamic feature according to a user's designation | designated.

  In addition, the acquisition unit 131 acquires a reaction of a user who has viewed the page generated by the generation unit 135. The generation unit 135 updates the generated page based on the reaction from the user.

  Thus, the extraction apparatus 100 according to the embodiment can perform the extraction process by feeding back the user's reaction. Thereby, the extraction device 100 can optimize the page to be generated.

  In addition, the acquisition unit 131 acquires information related to a user who has viewed the page generated by the generation unit 135. The generation unit 135 generates a page corresponding to the information related to the user, which is a page related to the predetermined product based on the information related to the user.

  Thus, the extraction device 100 according to the embodiment may generate a different page for each user. Thereby, since the extraction apparatus 100 can provide a page in which feature information according to the feature of each user is appealed, the appealing effect of the page can be improved.

  In addition, the extraction unit 134 extracts a phrase that characterizes the product as the feature information. The generation unit 135 generates a page in which the word / phrase extracted by the extraction unit 134 among the text data associated with the product is highlighted.

  As described above, the extraction apparatus 100 according to the embodiment may generate a page in such a manner that only features are highlighted with respect to the original explanatory text. Thereby, the extraction apparatus 100 can appeal the feature that the said commercial material is excellent with respect to a set, without destroying layout and content, such as a display of the description of an original provider.

  Further, the generation unit 135 generates a summary sentence of text data associated with the product based on the feature information, and also generates a page to which the summary sentence is given.

  As described above, the extraction apparatus 100 according to the embodiment may generate a summary sentence that summarizes the features of the product as the feature information. As a result, the extraction apparatus 100 can present the features of the merchandise to the user with short sentences instead of redundant explanations, and thus can generate a page that can efficiently convey information.

  Further, the generation unit 135 generates a page in which a summary sentence is arranged prior to the text data associated with the product.

  As described above, the extraction apparatus 100 according to the embodiment can quickly and accurately convey the feature of the product to the user by generating the page that presents the summary sentence prior to the original explanatory text. Thereby, the extraction apparatus 100 can improve the usability of the user who uses a commercial transaction site.

  In addition, the generation unit 135 generates a search result page including the feature information extracted by the extraction unit 134 when a product is displayed as a result of a predetermined search process.

  As described above, the extraction device 100 according to the embodiment may generate a page on which the feature information is presented as information displayed together with the search result. Thereby, the extraction apparatus 100 can convey the feature of the product to the user without the user bothering to access the appeal page.

  The embodiment of the present application has been described in detail with reference to the drawings. However, this is an exemplification, and various modifications and improvements are made based on the knowledge of those skilled in the art including the aspects described in the column of the disclosure of the invention. The present invention can be implemented in other forms.

  In addition, the “section (module, unit)” described above can be read as “means” or “circuit”. For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

DESCRIPTION OF SYMBOLS 1 Extraction processing system 10 User terminal 20 Provider terminal 30 Web server 100 Extraction apparatus 110 Communication part 120 Storage part 121 Commodity information storage part 122 Learning information storage part 123 Provision information storage part 130 Control part 131 Acquisition part 132 Learning part 133 Reception Part 134 extracting part 135 generating part 136 providing part

Claims (14)

An acquisition unit for acquiring product information associated with a product that is a target of a commercial transaction on the network;
For a predetermined product included in a set formed based on contents specified by the user, feature information that is information satisfying a predetermined condition in the set is extracted from the product information of the predetermined product. An extractor;
An extraction device comprising: The extraction unit includes:
When the product information of a predetermined product included in the set is information indicating superiority in the set, the product information is extracted as the feature information;
The extraction device according to claim 1. The extraction unit includes:
Extracting the feature information based on the predetermined condition set in advance for each category of the product;
The extraction apparatus according to claim 1 or 2, wherein The extraction unit includes:
The feature information is extracted based on a learning process using correct data set for text data associated with the product among the product information.
The extraction device according to any one of claims 1 to 3, wherein The extraction unit includes:
Extracting the feature information using a learning device learned to extract information characterizing text data associated with the product among the product information,
The extraction device according to any one of claims 1 to 4, wherein A generating unit that generates content related to the predetermined product based on the feature information extracted by the extracting unit;
The extraction apparatus according to claim 1, further comprising: The acquisition unit
Obtaining the reaction of the user who viewed the content generated by the generation unit,
The generator is
Updating the generated content based on the user's reaction;
The extraction apparatus according to claim 6. The acquisition unit
Obtain information about the user who viewed the content generated by the generation unit,
The generator is
Based on information related to the user, content related to the predetermined product, and content corresponding to the information related to the user is generated.
The extraction apparatus according to claim 6 or 7, characterized in that: The extraction unit includes:
As the feature information, a word that characterizes the product is extracted,
The generator is
Generating the content in a form in which the words extracted by the extraction unit are highlighted among the text data associated with the product;
The extraction apparatus according to any one of claims 6 to 8, wherein The generator is
Based on the feature information, a summary sentence of text data associated with the product is generated, and the content to which the summary sentence is attached is generated.
The extraction device according to any one of claims 6 to 9, wherein The generator is
Prior to the text data associated with the product, the content of the aspect in which the summary sentence is arranged is generated.
The extraction device according to claim 10. The generator is
When the product is displayed as a result of a predetermined search process, a search result page including the feature information extracted by the extraction unit is generated as the content.
The extraction device according to any one of claims 6 to 11, wherein An extraction method performed by a computer,
An acquisition step of acquiring product information associated with a product that is a target of a commercial transaction on the network;
For a predetermined product included in a set formed based on contents specified by the user, feature information that is information satisfying a predetermined condition in the set is extracted from the product information of the predetermined product. An extraction process;
The extraction method characterized by including. An acquisition procedure for acquiring merchandise information associated with a merchandise subject to commercial transactions on the network;
For a predetermined product included in a set formed based on contents specified by the user, feature information that is information satisfying a predetermined condition in the set is extracted from the product information of the predetermined product. Extraction procedure;
An extraction program characterized by causing a computer to execute.

Images