Recommendation system and method for support project based on user behavior pattern using deep learning

Abstract

The technology disclosed in this specification relates to a support project recommendation system based on user behavior patterns using deep learning, which recommends customized government support projects to companies by utilizing user behavior patterns and deep learning technology. The system includes an information providing unit (100) that provides a plurality of support project information (1), a class classification unit (200) that classifies the plurality of support project information (1) into a plurality of groups based on a pattern selected by a user terminal (10) connected to the information providing unit (100), a pattern matching unit (300) that matches the selected support project information (1) to the classified class based on the support project information (1) selected by another user terminal (10) connected to the information providing unit (100), and an information recommendation unit (400) that recommends support project information (1) according to the pattern of the matched group to the matched user terminal (10).

Description

Technology Field

[0001] The technology disclosed in this specification relates to a system that provides specific information tailored to a user, and is a support project recommendation system based on user behavior patterns using deep learning that recommends customized government support projects to companies by utilizing user behavior patterns and deep learning technology. Background Technology

[0002] Most conventional recommendation systems are filtering methods or rule-based recommendation systems that recommend suitable content to a user based on the interests the user has set.

[0003] Conventional Korean Patent Publication No. 10-2198148 (published Jan. 4, 2021, "System for Providing User Behavior Pattern Recommendation Service for Holistic Education") discloses a technology comprising a profile database storing user-specific profiles, a behavior pattern database storing multiple behavior pattern information, and a behavior pattern recommendation unit that recommends user behavior patterns in a specific space provided with multiple different behavior areas, while recommending behavior patterns by considering the user profile.

[0004] Conventional Korean Registered Patent Publication No. 10-1508330 (published on April 9, 2015, "Method for Providing Exercise Consulting Information") discloses a technology that can determine in real time whether the exercise consulting information provided to a user is suitable for the user by considering the user's exercise immersion and goal achievement value for each unit monitoring cycle within a target period composed of multiple unit monitoring cycles, and can automatically update the exercise consulting information by learning to match the exercise consulting information most preferred by the user and having exercise effects based on the exercise consulting information determined in the final unit monitoring cycle or the exercise consulting information determined for each unit monitoring cycle to the user's information.

[0005] Conventional Korean Patent Publication No. 10-2012-0075537 (published July 9, 2012, "System and Method for Diagnosing Corporate Competitiveness") includes: a database storing data on multiple diagnostic items for diagnosing the competitiveness of a target company and data on support projects being conducted to support the companies; a diagnostic data collection module that generates a web page composed of the diagnostic item data, displays it on a terminal of the target company, and collects diagnostic data entered into the web page; a competitiveness diagnostic module that diagnoses the competitiveness of the target company according to the diagnostic items using the diagnostic data and derives a diagnostic result; a support project linkage module that searches for support projects recommendable to the target company from the support project data based on the diagnostic result; and a reporting module that displays the diagnostic result of the competitiveness diagnostic module and the search result of the support project linkage module on the terminal of the target company. Accordingly, a technology is disclosed that enables companies to self-diagnose their competitiveness using the answers to diagnostic items, while also suggesting additional or supplementary measures based on the diagnostic results and recommending corporate support programs suitable for the diagnostic results. The problem to be solved

[0006] The technology disclosed in this specification is a technology for recommending government support programs tailored to companies. It aims to solve the problem of providing a support program recommendation system based on user behavior patterns using deep learning, which analyzes and classifies the behavior patterns (log) of users with similar corporate characteristics to form groups, and recommends support programs preferred by similar users. means of solving the problem

[0007] In one embodiment, a support project recommendation system based on user behavior patterns using deep learning is disclosed.

[0008] The support business recommendation system based on user behavior patterns using deep learning disclosed in this specification includes an information providing unit (100) that provides a plurality of support business information (1), a class classification unit (200) that classifies the plurality of support business information (1) into a plurality of groups based on a pattern selected by a user terminal (10) connected to the information providing unit (100), a pattern matching unit (300) that matches the classified class based on the support business information (1) selected by another user terminal (10) connected to the information providing unit (100), and an information recommendation unit (400) that recommends support business information (1) according to the pattern of the matched group to the matched user terminal (10).

[0009] The class classification unit (200) may include a pattern generation unit (210) that generates pattern information (2) by recording support business information (1) selected by a user terminal (10) connected to the information provision unit (100), and a pattern grouping unit (220) that classifies various patterns according to the pattern information (2).

[0010] The support business recommendation system based on user behavior patterns using deep learning disclosed in this specification may include linking the information provider (100) and the chatbot.

[0011] The support business recommendation system based on user behavior patterns using deep learning disclosed in this specification further includes a personal information collection unit that collects personal information of a user, and the class classification unit (200) may include creating and classifying class groups through personal information and support business information (1) selected from a user terminal (10). Effects of the invention

[0012] The support project recommendation system based on user behavior patterns using deep learning disclosed in this specification has the effect that when a user clicks and selects a support project information (1) of interest among a plurality of support project information (1) provided by an information provider (100) using a user terminal (10), other support project information (1) selected by other users who selected the same support project information (1) as the support project information (1) selected by the user can be provided by the information recommendation unit (400) through the pattern matching unit (300).

[0013] The support project recommendation system based on user behavior patterns using deep learning disclosed in this specification has the effect of enabling smooth recommendation of support project information (1) through artificial intelligence by patterning the support project information (1) selected by users through the class classification unit (200) and classifying them into class groups.

[0014] The foregoing provides only optional concepts in a simplified form regarding matters to be described in more detail later. This content is not intended to limit the main or essential features of the patent claims, or to restrict the scope of the patent claims. Brief explanation of the drawing

[0015] FIG. 1 is a diagram roughly illustrating the configuration of a support business recommendation system based on user behavior patterns using deep learning disclosed in this specification. FIG. 2 is a diagram illustrating the operational relationship of a support business recommendation system based on user behavior patterns using deep learning disclosed in this specification. FIG. 3 is a diagram roughly illustrating a support project recommendation method based on user behavior patterns using deep learning disclosed in this specification. FIG. 4 is a drawing illustrating types classified by the class classification unit disclosed in this specification. Specific details for implementing the invention

[0016] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the drawings. Unless otherwise specified in the text, similar reference numerals in the drawings indicate similar components. The exemplary embodiments described in the detailed description, drawings, and claims are not intended to be limiting, and other embodiments may be used, and other modifications are possible without departing from the spirit or scope of the art disclosed herein. A person skilled in the art will readily understand that the components of this disclosure—namely, those generally described herein and those described in the drawings—can be arranged, configured, combined, or designed in various different configurations, all of which are clearly devised and form part of this disclosure. In the drawings, the width, length, thickness, or shape of components may be exaggerated to clearly represent various layers (or films), regions, and shapes.

[0017] Where one component is referred to as being "provided" to another component, this may include cases where the one component is directly provided to the other component, as well as cases where an additional component is interposed between them.

[0018] When one component is referred to as being "provided" to another component, it may include cases where the one component is directly provided to the other component, as well as cases where an additional component is interposed between them.

[0019] Since the description of the disclosed technology is merely an example for structural or functional explanation, the scope of rights of the disclosed technology should not be interpreted as being limited by the examples described in the text. That is, since the examples can be modified in various ways and can take various forms, the scope of rights of the disclosed technology should be understood to include equivalents capable of realizing the technical concept.

[0020] A singular expression should be understood to include a plural expression unless the context clearly indicates otherwise, and terms such as “include” or “have” are intended to specify the existence of the implemented features, numbers, steps, actions, components, parts, or combinations thereof, and should not be understood as precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0021] Unless otherwise defined, all terms used herein have the same meaning as generally understood by a person skilled in the art to which the disclosed technology belongs. Terms defined in commonly used dictionaries should be interpreted as having meanings consistent with the context of the relevant technology and should not be interpreted as having an ideal or overly formal meaning unless explicitly defined in this application.

[0022] The support business recommendation system based on user behavior patterns using deep learning disclosed in this specification comprises an information providing unit (100) that provides a plurality of support business information (1) approximately; a class classification unit (200) that classifies the plurality of support business information (1) into a plurality of groups based on a pattern selected by a user terminal (10) connected to the information providing unit (100); a pattern matching unit (300) that matches the classified class based on the support business information (1) selected by another user terminal (10) connected to the information providing unit (100); and an information recommendation unit (400) that recommends support business information (1) according to the pattern of the matched group to the matched user terminal (10).

[0023] The class classification unit (200) may include a pattern generation unit (210) that generates pattern information (2) by recording support business information (1) selected by a user terminal (10) connected to the information provision unit (100), and a pattern grouping unit (220) that classifies various patterns according to the pattern information (2).

[0024] The support business recommendation system based on user behavior patterns using deep learning disclosed in this specification may optionally further include linking the information provider (100) and the chatbot.

[0025] The following description is explained in more detail with reference to the drawings attached to this specification.

[0026] FIG. 1 of the drawings attached to this specification is a diagram roughly illustrating the configuration of a support project recommendation system based on user behavior patterns using deep learning disclosed in this specification. FIG. 2 is a diagram illustrating the operational relationship of a support project recommendation system based on user behavior patterns using deep learning disclosed in this specification. FIG. 3 is a diagram roughly illustrating a support project recommendation method based on user behavior patterns using deep learning disclosed in this specification.

[0027] According to one embodiment with reference to FIG. 1 or FIG. 2, the support business recommendation system based on user behavior patterns using deep learning disclosed in this specification includes an information providing unit (100), a class classification unit (200), a pattern matching unit (300), and an information recommendation unit (400).

[0028] The information providing unit (100) provides multiple support project information (1). The information providing unit (100) may be set up as a standard server. The information providing unit (100) may be linked with a standard webpage, a standard application, or a standard chatbot. The information providing unit (100) may be connected to a database in which various support project information (1) is stored or to an API in which the support project information (1) is stored. Support project information (1) refers to information regarding various government or private support projects that support companies or individuals. The information providing unit (100) may provide various support project information (1) as a list of announcements arranged in the order of 1, 2, 3, 4, 5, 6, etc. The information providing unit (100) is connected to a user terminal (10). The information providing unit (100) causes the support project information (1) to be output to the user terminal (10) through an input value selected by the user terminal (10) from among the listed support project information (1). That is, the information provider (100) provides support project information (1) to connected users.

[0029] Meanwhile, users can access the information provider (100) through a user terminal (10) such as a standard smartphone, computer, or PDA.

[0030] The class classification unit (200) classifies multiple support project information (1) into multiple groups based on a pattern selected by a user terminal (10) connected to the information provision unit (100). The class classification unit (200) can be provided on a server, etc., using conventional artificial intelligence. The artificial intelligence can be built through machine learning based on a conventional softmax regression method. That is, the class classification unit (200) classifies the support project information (1) selected by the user terminal (10) from among the support project information (1) listed through the information provision unit (100) by patterning it. The class classification unit (200) classifies patterns through past records.

[0031] The class classification unit (200) may include a pattern generation unit (210) and a pattern grouping unit (220). The class classification unit (200) may be connected to the information provision unit (100).

[0032] The pattern generation unit (210) generates pattern information (2) through the recording of support project information (1) selected by a user terminal (10) connected to the information provision unit (100). To explain in more detail, the pattern generation unit (210) can be connected to the information provision unit (100). The pattern generation unit (210) collects past input values ​​in which the user terminal (10) selected support project information (1). The pattern generation unit (210) generates pattern information (2) in which the collected support project information (1) selected by the user terminal (10) is formed into a single pattern. For example, referring to FIG. 4, if there is an input value in which a user terminal (10) selects support project information (1) number 1, 3, 5, and 7 from the listed support project information (1), the pattern generation unit (210) records support project information (1) number 1, 3, 5, and 7 as a single pattern information (2).

[0033] Meanwhile, the pattern generation unit (210) may record support business information (1) selected for each user through the personal information of the user using the user terminal (10).

[0034] The pattern grouping unit (220) classifies various patterns according to pattern information (2). The pattern grouping unit (220) compares the pattern information (2) generated by the pattern generation unit (210) to extract identical patterns. The pattern grouping unit (220) groups the extracted patterns. The pattern grouping unit (220) creates multiple classes by making the grouped patterns into one class as shown in FIG. 4. That is, the classification is determined so that numbers 1, 3, 5, and 7 of the support project information (1) form one pattern information (2) as Class 1. Ultimately, the pattern grouping unit (220) groups various pattern information (2) together to classify them into classes.

[0035] Meanwhile, the pattern grouping unit (220) may limit the number of support business information (1) selected by the user terminal (10) and set various classes such as the number of classes generated by comparing pattern information (2) with only the limited number.

[0036] The pattern matching unit (300) matches the support project information (1) selected by another user terminal (10) connected to the information providing unit (100) to a classified class. To explain in more detail, the pattern matching unit (300) collects input values ​​for selecting support project information (1) from the user terminal (10) in the information providing unit (100). The pattern matching unit (300) extracts class groups based on pattern information (2) containing support project information (1) corresponding to the collected input values ​​for selecting support project information (1). Here, the more support project information (1) selected by the user terminal (10), the higher the matching accuracy of the class groups.

[0037] As an example with reference to FIG. 4, when a user terminal (10) selects support project information (1) 3 and 5, a Class 1 group including pattern information (2) containing support project information (1) 3 and 5 is matched.

[0038] The information recommendation unit (400) recommends support business information (1) according to the pattern of the group matched to the user terminal (10). To explain in more detail, the information recommendation unit (400) recommends support business information (1) included in the pattern information (2) of the class group matched by the pattern matching unit (300).

[0039] As an example with reference to FIG. 4, when a user terminal (10) selects support business information (1) 3 and 5, it recommends support business information (1) 1 and 7 of pattern information (2) included in class 1 group.

[0040] The support project recommendation system based on user behavior patterns using deep learning described above has the effect that when a user clicks and selects a support project information (1) of interest from among the multiple support project information (1) provided by the information provider (100) using a user terminal (10), other support project information (1) selected by other users who selected the same support project information (1) as the support project information (1) selected by the user can be provided by the information recommendation unit (400) through the pattern matching unit (300).

[0041] Meanwhile, by patterning the support project information (1) selected by users through the class classification unit (200) and classifying them into class groups, there is an effect that the recommendation of support project information (1) through artificial intelligence can be smoothly carried out.

[0042] According to another embodiment, the support project recommendation system based on user behavior patterns using deep learning disclosed in this specification may include linking an information provider (100) with a chatbot. The chatbot may utilize a standard Open Kakao chatbot, etc. Input values ​​and support project information (1) entered from a user terminal (10) can be collected and provided through the chatbot.

[0043] According to another embodiment, the support project recommendation system based on user behavior patterns using deep learning disclosed in this specification may further include a personal information collection unit (not shown). The personal information collection unit requests information input so that user personal information can be entered through a user terminal (10). The personal information may consist of an address or region, areas of interest, and business type. Areas of interest may be classified into fields such as technology, startup, or loan of government support projects. Business types may be classified into individuals, sole proprietors, corporate businesses, etc. The personal information collection unit may include personal information in the pattern information (2). At this time, the class classification unit (200) creates class groups and classifies them through the personal information and support project information (1) selected from the user terminal (10). The pattern matching unit (300) and the information recommendation unit (400) operate through the classified class groups. Ultimately, the user is matched based on the selected patterns of personal information and government support project information (1).

[0044] Meanwhile, a server running the chatbot may also be constructed to build a support business recommendation system based on user behavior patterns using deep learning disclosed in this specification.

[0045] The support project recommendation method based on user behavior patterns using deep learning with reference to Fig. 3 largely includes an information provision process (S100), a class classification process (S200), a pattern matching process (S300), and an information recommendation process (S400).

[0046] The information provision process (S100) provides multiple support project information (1) to a user terminal (10). The information provision process (S100) provides the support project information (1) to the user terminal (10) through a webpage or application, etc., in the form of a list of announcements arranged in the order of 1, 2, 3, 4, 5, 6, ? through a database in which various support project information (1) is stored or through an API in which the support project information (1) is stored. The information provision process (S100) outputs the details of the support project information (1) to the user terminal (10) through an input value selected by the user terminal (10) from among the listed support project information (1).

[0047] The class classification process (S200) classifies multiple support project information (1) into multiple class groups based on a pattern selected by a user terminal (10) connected to the information provision unit (100). The class classification process (S200) patterns the support project information (1) selected by the user terminal (10) from among the support project information (1) listed through the information provision unit (100). The class classification process (S200) classifies class groups based on the pattern. The class classification process (S200) classifies the pattern through past records.

[0048] The pattern matching process (S300) matches the support project information (1) selected by another user terminal (10) connected to the information providing unit (100) to a class group classified based on the information. More specifically, the pattern matching process (S300) collects input values ​​for selecting support project information (1) from the user terminal (10) in the information providing unit (100). The pattern matching process (S300) extracts a class group based on pattern information (2) containing support project information (1) corresponding to the collected input values ​​for selecting support project information (1).

[0049] The information recommendation process (S400) recommends support project information (1) according to the pattern of the group matched to the user terminal (10). To explain in more detail, the information recommendation process (S400) recommends support project information (1) included in the pattern information (2) of the class group matched in the pattern matching unit (300).

[0050] From the foregoing, it will be understood that various embodiments of the present disclosure have been described for illustrative purposes, and that various possible modifications exist without departing from the scope and spirit of the present disclosure. Furthermore, the various embodiments disclosed are not intended to limit the spirit of the present disclosure, and the true spirit and scope will be set forth in the following claims. Explanation of the symbols

[0051] 1: Support Program Information 2 : Pattern Information 10 : User terminal 100 : Information provider 200 : Class classification section 210 : Pattern generation unit 220 : Pattern Grouping Section 300 : Pattern matching section 400 : Information Recommendation Section S100: Information Provision Process S200: Class classification process S300: Pattern Matching Process S400: Information Recommendation Process

Claims

Claim 1 A support project recommendation system based on user behavior patterns using deep learning, comprising: an information provision unit that provides multiple support project information; a class classification unit that classifies multiple support project information into multiple groups based on a pattern of selection by a user terminal connected to the information provision unit; a pattern matching unit that matches the classified classes based on support project information selected by another user terminal connected to the information provision unit; and an information recommendation unit that recommends support project information according to the pattern of the matched group to the matched user terminal. Claim 2 In claim 1, the class classification unit comprises: a pattern generation unit that generates pattern information by recording support business information selected by a user terminal connected to the information provision unit; and a pattern grouping unit that classifies into various patterns according to the pattern information; a support business recommendation system based on user behavior patterns using deep learning. Claim 3 In paragraph 2, the information providing unit is a support project recommendation system based on user behavior patterns using deep learning, which includes linking with a chatbot.

Images