KR102652416B1 - A method for presenting information on a product based on high speed similarity calculation and an apparatus therefor - Google Patents

Abstract

The present invention relates to a method for providing information about products based on high-speed similarity calculation and a recommendation server for the same. The present invention relates to a method for providing information on products based on high-speed similarity calculation, and a recommendation server for the same. This invention relates to a data collection server collecting a plurality of behavior logs, each containing a user identifier and a product, from a user terminal. Collecting and transmitting to an analysis server, the analysis server deleting some of the plurality of behavior logs, and the analysis server performing a similarity calculation on the plurality of partially erased behavior logs to calculate the similarity. Information about the product comprising a step of the analysis server generating information about the product according to the similarity calculation result, and the step of the analysis server transmitting the information about the product to an application server. Provides a method for providing and a recommendation server for this.

Description

A method for presenting information on a product based on high speed similarity calculation and an apparatus therefor}

The present invention relates to a technology for providing information about a product. More specifically, the present invention relates to a technology for providing information about a product, and more specifically, provides information about the product through similarity analysis of behavior logs, while reducing the amount of data for similarity calculation to enable high-speed similarity calculation. It relates to a method and device for deriving information about a product and providing the derived information.

As the application market running on mobile devices such as smartphones and tablet pads is rapidly expanding, the related advertising market is also expanding at a rapid pace.

In this regard, mobile devices (hereinafter referred to as “user terminals”) may receive and display advertisements provided from a recommendation server when running an application that is an advertising medium or accessing a service page (e.g., a web page), and the displayed advertisements If you select , you can access the advertiser's page linked to the advertisement.

At this time, in order to increase advertising effectiveness, the recommendation server verifies the properties of the advertisement registered by the advertiser and the appropriate advertisement recipient group, and adopts a type of targeting technique that provides the advertisement to each user belonging to the confirmed advertisement recipient group.

Here, the advertisement recipient group can be classified by a segmentation technique that defines the attributes corresponding to each user's advertisement as a segment and classifies users corresponding to the same segment into one advertisement recipient group. A group of advertisement recipients classified by defining specific attributes according to arbitrary classification criteria can be referred to as a segment.

In existing targeting advertisements, each segment is extracted based on the user's behavior log information, and the segment is extracted by analyzing the similarity of the behavior log information. However, the amount of behavior log information is so large that the computation time for similarity analysis places a huge load on the system. Therefore, a method that can reduce the calculation time is required.

Korean Patent No. 10-1567217, registered on November 2, 2015 (Name: Collaborative filtering recommendation method and system using social network analysis techniques)

Therefore, the purpose of the present invention is to extract information about products through high-speed similarity calculation by reducing the amount of data for similarity calculation when extracting information about products through similarity analysis of a large amount of behavior log information. The aim is to provide a method for providing extracted information and a device for this.

In order to achieve the above-described object, a method for providing information about products according to a preferred embodiment of the present invention involves the data collection server collecting a plurality of behavior logs, each containing a user identifier and product, from the user terminal. Transmitting to an analysis server, the analysis server deleting some of the plurality of behavior logs, and the analysis server performing a similarity calculation on the plurality of partially erased behavior logs to calculate similarity; It includes the step of the analysis server generating information about the product according to the similarity calculation result, and the step of the analysis server transmitting the information about the product to the application server.

The step of erasing some of the plurality of behavior logs includes determining an erasure ratio, which is the ratio of behavior logs to be erased from the entire behavior log, depending on the amount of information about the product to be provided.

The step of erasing some of the plurality of behavior logs is characterized in that, after the step of determining the erasing rate, the number of purchased products of the behavior logs to be erased is determined according to the determined erasing rate.

The step of calculating the similarity by performing the similarity calculation uses as the denominator the union of action logs that include at least one of one product and another product, and the intersection of action logs that includes both the one product and the other product. It is characterized by calculating the similarity between one product and the other product using the numerator.

The step of generating information about a product according to the similarity calculation result is characterized by deriving a predetermined number of products as information about the product in the order of the highest similarity according to the calculated similarity.

In order to achieve the above-described object, a device for providing information about products according to a preferred embodiment of the present invention includes an analysis server communication unit for communication with an application server, and a target object from the application server through the analysis server communication unit. Upon receiving a recommended product request message containing the product and the number of recommended products, the deletion target is specified among a plurality of behavior logs including a user identifier and a purchased product corresponding to the user identifier, and the specified deletion target is selected from the plurality of behavior logs. It includes a preprocessing module that erases the object to be erased, and a similarity calculation module that calculates the degree of similarity with the target product using a plurality of behavior logs from which the object to be erased has been erased.

The preprocessing module calculates an erasure rate, which is the ratio of behavior logs that can be erased according to the number of recommended products, and determines an erasure target according to the calculated erasure rate and the number of purchased products in the behavior log. .

The similarity calculation module uses as the denominator the union of action logs including at least one of the target product and the other product, and as the numerator the intersection of action logs including both the target product and the other product, and calculates the difference between the target product and the other product. It is characterized by calculating similarity.

The device derives the number of recommended products as recommended products in the order of the highest similarity among the other products according to the calculated similarity, and transmits the derived recommended products to the application server through the analysis server communication unit. It further includes an information provision module.

According to the present invention, the speed of similarity calculation is improved by specifying a plurality of behavior logs that do not affect the result of similarity calculation as objects of erasure and erasing them in the pre-processing process, thereby reducing the load on the system.

1 is a schematic configuration diagram of an information provision system according to an embodiment of the present invention.
Figure 2 is a block diagram showing the main configuration of a user terminal according to an embodiment of the present invention.
Figure 3 is a block diagram showing the main configuration of a data collection server according to an embodiment of the present invention.
Figure 4 is a block diagram showing the main configuration of an analysis server according to an embodiment of the present invention.
Figure 5 is a block diagram showing the main configuration of an application server according to an embodiment of the present invention.
Figure 6 is a flowchart illustrating a method for providing information about a product according to an embodiment of the present invention.
Figure 7 is a flowchart illustrating a high-speed similarity calculation method according to an embodiment of the present invention.
Figure 8 is a flowchart illustrating a method of providing information about a product according to another embodiment of the present invention.
Figure 9 is an example screen for explaining a method of providing information about a product according to another embodiment of the present invention.

In order to make the characteristics and advantages of the problem-solving means of the present invention clearer, the present invention will be described in more detail with reference to specific embodiments of the present invention shown in the attached drawings.

However, detailed descriptions of known functions or configurations that may obscure the gist of the present invention are omitted in the following description and attached drawings. Additionally, it should be noted that the same components throughout the drawings are indicated by the same reference numerals whenever possible.

Terms or words used in the following description and drawings should not be construed as limited to their usual or dictionary meanings, and the inventor may appropriately define the concept of the term to explain his or her invention in the best way. It must be interpreted with meaning and concept consistent with the technical idea of the present invention based on the principle that it is. Therefore, the embodiments described in this specification and the configuration shown in the drawings are only one of the most preferred embodiments of the present invention, and do not represent the entire technical idea of the present invention, so at the time of filing the present application, various alternatives may be used to replace them. It should be understood that equivalents and variations may exist.

In addition, terms containing ordinal numbers, such as first, second, etc., are used to describe various components, and are used only for the purpose of distinguishing one component from other components and to limit the components. Not used. For example, the second component may be referred to as the first component without departing from the scope of the present invention, and similarly, the first component may also be referred to as the second component.

In addition, when a component is referred to as being “connected” or “connected” to another component, it means that it is or can be connected logically or physically. In other words, a component may be directly connected or connected to another component, but it should be understood that other components may exist in the middle and may be indirectly connected or connected.

Additionally, the terms used in this specification are only used to describe specific embodiments and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In addition, terms such as "comprise" or "have" used in the specification are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are not intended to indicate the presence of one or more of the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification. It should be understood that this does not exclude in advance the presence or addition of other features, numbers, steps, operations, components, parts, or combinations thereof.

In addition, terms such as "unit", "unit", and "module" used in the specification refer to a unit that processes at least one function or operation, which is implemented as hardware, software, or a combination of hardware and software. It can be.

Additionally, the words “a or an,” “one,” “the,” and similar words may indicate otherwise herein in the context of describing the invention (particularly in the context of the claims below). It may be used to include both singular and plural terms, unless clearly contradicted by the context.

Additionally, embodiments within the scope of the present invention include computer-readable media having or transmitting computer-executable instructions or data structures stored on the computer-readable media. Such computer-readable media may be any available media that can be accessed by a general-purpose or special-purpose computer system. By way of example, such computer-readable media may include RAM, ROM, EPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage, or in the form of computer-executable instructions, computer-readable instructions or data structures. It may be used to store or transmit certain program code means, and may include, but is not limited to, a physical storage medium such as any other medium that can be accessed by a general purpose or special purpose computer system. .

In the following description and claims, “network” is defined as one or more data links that enable the transfer of electronic data between computer systems and/or modules. When information is transmitted or provided to a computer system over a network or other communication connection (wired, wireless, or a combination of wired and wireless), this connection may be understood as a computer-readable medium. Computer-readable instructions include, for example, instructions and data that cause a general-purpose or special-purpose computer system to perform a particular function or group of functions. Computer-executable instructions may be binary, intermediate format instructions, for example, assembly language, or even source code.

In addition, the present invention is applicable to personal computers, laptop computers, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile phones, PDAs, and pagers. In a network computing environment with various types of computer system configurations including pagers, etc., it can be applied to providing advertisements to the computer systems. The invention may also be practiced in distributed system environments where tasks are performed by both local and remote computer systems that are linked over a network via wired data links, wireless data links, or a combination of wired and wireless data links. In a distributed system environment, program modules may be located in local and remote memory storage devices.

First, an information provision system according to an embodiment of the present invention will be described. 1 is a schematic configuration diagram of an information provision system according to an embodiment of the present invention.

Referring to Figure 1, the information provision system 100 according to an embodiment of the present invention includes a plurality of user terminals (10: 10-1, 10-2,??10-n), a data collection server 20, and analysis. Includes a server 30 and an application server 40.

A plurality of user terminals (10: 10-1, 10-2,??10-n), data collection server 20, analysis server 30, and application server 40 exchange data with each other through a communication network (NW). can be sent and received.

For example, the communication network (NW) may include wireless communication networks such as Wireless LAN (WLAN), Wi-Fi, Wibro, Wimax, and High Speed Downlink Packet Access (HSDPA), Depending on how the system is implemented, it may include wired communication networks such as Ethernet, xDSL (ADSL, VDSL), HFC (Hybrid Fiber Coaxial Cable), FTTC (Fiber to The Curb), and FTTH (Fiber To The Home).

In addition, the communication network (NW) of the present invention may include, for example, a mobile communication network consisting of a plurality of access networks (not shown) and a core network (not shown) connecting them. Here, the access network is a network that performs wireless communication by directly connecting to the terminal, for example, a number of base stations such as BS (Base Station), BTS (Base Transceiver Station), NodeB, eNodeB, etc., and BSC (Base Station Controller) , can be implemented with a base station controller such as RNC (Radio Network Controller). In addition, as described above, the digital signal processing unit and the wireless signal processing unit integrated in the base station are divided into a digital unit (hereinafter referred to as DU) and a radio unit (hereinafter referred to as RU), respectively. It can also be configured by installing a number of RUs (not shown) in multiple areas and connecting the multiple RUs to a centralized DU. In addition, the core network (not shown) that forms the mobile network together with the access network is connected to the access network. It serves to connect other communication networks such as external networks, for example, the Internet network.

As explained above, this core network is a network system that performs major functions for mobile communication services such as mobility control and switching between access networks, and performs circuit switching or packet switching, and is a mobile network system that performs circuit switching or packet switching. Manage and control packet flow within. In addition, the core network may manage inter-frequency mobility and play a role in interoperating with traffic within the access network and core network and with other networks, such as the Internet network. This core network will further include Serving GateWay (SGW), PDN GateWay (PGW), Mobile Switching Center (MSC), Home Location Register (HLR), Mobile Mobility Entity (MME), and Home Subscriber Server (HSS). It may be possible.

Additionally, the communication network (NW) according to the present invention may include an Internet network. The Internet network refers to a normal open communication network, that is, a public network, in which information is exchanged according to the TCP/IP protocol.

Through this communication network (NW), multiple user terminals (10: 10-1, 10-2,??10-n), data collection server 20, analysis server 30, and application server 40 are interconnected. Thus, an information provision system 100 for providing product information according to the present invention is configured.

Each of the plurality of user terminals 10 can install and execute at least one application. In particular, the user terminal 10 may install an application that interoperates with the data collection server 20 and transmit a behavior log to the data collection server 20 according to execution of the application. At this time, the plurality of user terminals 10 can transmit the behavior log to the data collection server 20 through an installed SDK (Software Development Kit). However, it is not limited to this, and the plurality of user terminals 10 may provide a behavior log to the data collection server 20 through installation of a separate application in conjunction with the data collection server 20.

At this time, the plurality of user terminals 10 may extract a behavior log containing product purchase history through an API (Application Programming Interface) included in the installed SDK, and provide the extracted behavior log to the data collection server 20. . Here, the behavior log includes a user identifier (UID: User ID), which is a unique code of the user terminal, and purchased products (items). Here, the user identifier (UID) may be an identification number assigned to each mobile terminal in order to distinguish each user terminal 10.

Here, the plurality of user terminals 10 are described as mobile communication terminals as a representative example, but the terminal is not limited to mobile communication terminals, and includes all information and communication devices, multimedia terminals, wired terminals, fixed terminals, and IP (Internet Protocol) terminals. It can be applied to various terminals. In addition, the terminal can be advantageously used when it is a mobile terminal with various mobile communication specifications such as a mobile phone, PMP (Portable Multimedia Player), MID (Mobile Internet Device), smart phone, and information and communication devices. .

The data collection server 20 collects behavior logs including user identifiers (UID) and products purchased by users from a plurality of user terminals 10, and delivers the collected behavior logs to the analysis server 30. You can.

The analysis server 30 can sort the received behavior logs according to user identifier (UID) and product (item). For example, the analysis server 30 can create a table with user identifiers (UID) as columns and products (items) as rows. And the analysis server 30 can analyze the behavior log collected from the data collection server 20 and derive information related to the product. For example, the analysis server 30 can analyze the behavior log to derive 'products purchased together'. In other words, the analysis server 30 can derive the target product (TG) and the 'product purchased together (RG)' through similarity calculation. The analysis server 30 provides information about the derived product, for example, 'product purchased together (RG)' to the application server 40.

The application server 40 uses information about these products to provide additional services for product marketing. Therefore, the application server 40 can provide information about the product when there is a request from the user terminal 10.

Meanwhile, in the embodiment of the present invention, the data collection server 20, the analysis server 30, and the application server 40 are described as independent devices, but the shopping server 40 and the recommendation server 50 are not limited thereto. ) may be implemented as a single device.

Next, the user terminal 10 according to an embodiment of the present invention will be described in more detail. Figure 2 is a block diagram showing the main configuration of a user terminal according to an embodiment of the present invention.

Referring to FIG. 2, the user terminal 10 according to an embodiment of the present invention includes a communication unit 11, an input unit 13, a display unit 15, a storage unit 17, and a control unit 19.

The communication unit 11 functions to perform communication of the user terminal 10. That is, the communication unit 11 transmits and receives data to and from the data collection server 20 through the communication network (NW). In particular, the communication unit 11 according to an embodiment of the present invention can transmit a behavior log to the data collection server 20.

The input unit 13 receives various information such as number and character information, and transmits signals input related to setting various functions and controlling functions of the user terminal 10 to the control unit 19. Additionally, the input unit 13 may be configured to include at least one of a keypad and a touchpad that generate an input signal according to a user's touch or manipulation. At this time, the input unit 13 is configured in the form of a touch panel (or touch screen) together with the display unit 15 and can perform input and display functions simultaneously. In particular, the input unit 13 can transmit input signals for communication and data according to application execution to the control unit 19 according to the user's manipulation. Additionally, when the user terminal 10 is configured as a touch screen, the input unit 13 can receive signals touched on the screen or inputs such as a keyboard and mouse.

The display unit 15 displays information about a series of operation states and operation results that occur while performing the functions of the user terminal 10. Additionally, the display unit 15 can display menus of the user terminal 10 and user data input by the user. Here, the display unit 15 includes LCD (Liquid Crystal Display), TFT-LCD (Thin Film Transistor LCD), OLED (Organic Light Emitting Diodes), light emitting diode (LED), AMOLED (Active Matrix Organic LED), and flexible display (Flexible Display). It can be composed of a display) and a 3-dimensional display (3 Dimension). In particular, when one of the plurality of web pages constituting the shopping mall website is received, the display unit 15 displays the received web page on the screen.

The storage unit 17 stores application programs necessary for functional operation of the user terminal 10. Here, when the user terminal 10 activates each function in response to the user's request, the user terminal 10 executes the corresponding application programs under the control of the control unit 19 to provide each function. Additionally, the data area is an area where data generated according to the use of the user terminal 10 is stored.

The control unit 19 can initialize each component of the user terminal 10 and perform necessary signal control. In particular, the control unit 19 according to an embodiment of the present invention connects to the shopping mall web server through the communication unit 11, receives web pages constituting the shopping mall website, and displays the received web pages on the screen through the display unit 15. It is displayed as Accordingly, the user can view the web page and purchase a certain product through the web page. In particular, when the control unit 19 detects a signal for selecting a specific product for purchase through the input unit 13 according to the user's operation, it transmits a purchase request for the corresponding product to the shopping mall web server through the communication unit 11. You can purchase the product. Then, the control unit 19 creates a behavior log through the purchase history of purchasing the product. Here, the behavior log includes a user identifier (UID) and purchased product (item). Subsequently, the control unit 19 may transmit the behavior log to the data collection server 20 through the communication unit 11.

Next, the configuration of the data collection server 20 according to an embodiment of the present invention will be described in more detail. Figure 3 is a block diagram showing the main configuration of a data collection server according to an embodiment of the present invention.

Referring to FIG. 3, the data collection server 20 according to an embodiment of the present invention includes a collection server communication unit 21, a collection server storage unit 23, and a collection server control unit 25.

The collection server communication unit 41 has an interface for communicating with the user terminal 10, analysis server 30, and application server 40 through a communication network (NW). In particular, the collection server communication unit 41 can receive a behavior log from the user terminal 10.

The collection server storage unit 23 stores data or programs executed or processed by the collection server control unit 25. This collection server storage unit 23 can store a behavior log mapping the user identifier (UID) received from the user terminal 10 and the purchased product temporarily or for a predetermined period of time.

The collection server control unit 25 serves to overall control the operation of the data collection server 20, including the collection server communication unit 21 and the collection server storage unit 23, and may include a processor for this purpose. This collection server control unit 25 stores the behavior log received from the user terminal 10 in the collection server storage unit 23. Additionally, the collection server control unit 25 may periodically transmit the behavior log stored in the collection server storage unit 23 to the analysis server 30 through the collection server communication unit 41.

Next, with reference to the drawings, the configuration of the analysis server 30 according to an embodiment of the present invention will be described in more detail. Figure 4 is a block diagram showing the main configuration of an analysis server according to an embodiment of the present invention. Referring to FIG. 4, the analysis server 30 according to an embodiment of the present invention includes an analysis server communication unit 31, an analysis server storage unit 33, and an analysis server control unit 35.

The analysis server communication unit 31 has an interface for communicating with the data collection server 20 and the application server 30 through a communication network (NW). In particular, when the analysis server communication unit 31 receives a behavior log from the data collection server 20, it transmits it to the analysis server control unit 35. In addition, the analysis server communication unit 31 receives product information (eg, a list of products purchased together) from the analysis server control unit 35 and transmits it to the application server 40.

The analysis server storage unit 33 is configured to store data or programs executed or processed by the analysis server control unit 35. This analysis server storage unit 33 can store programs that derive recommended products based on similarity from behavior logs. Additionally, the analysis server storage unit 33 can accumulate and store the behavior logs received from the data collection server 20.

The analysis server control unit 35 serves to overall control the operation of the analysis server 30, including the analysis server communication unit 31 and the analysis server storage unit 33, and may include a processor for this purpose.

In particular, the analysis server control unit 35 may be configured to include a preprocessing module 35a, a similarity calculation module 35b, and an information provision module 35c.

When the analysis server control unit 35 receives the behavior log through the analysis server communication unit 31, it stores it in the analysis server storage unit 32. The preprocessing module 35a loads a plurality of behavior logs stored in the analysis server storage unit 32 and erases part of the behavior logs. At this time, the preprocessing module 55a erases some of the plurality of behavior logs. The similarity calculation module 35b calculates the similarity between the plurality of behavior logs, some of which have been erased. This similarity can be calculated using the Jacquard similarity calculation technique. The information provision module 55c derives product information, for example, recommended products such as 'products purchased together', through the similarity calculation results, and sends the derived recommended products to the application server 40 through the analysis server communication unit 51. ) can be transmitted.

Next, the configuration of the application server 40 according to an embodiment of the present invention will be described in more detail. Figure 5 is a block diagram showing the main configuration of an application server according to an embodiment of the present invention.

Referring to Figure 5, the application server 40 according to an embodiment of the present invention includes an application server communication unit 41, an application server storage unit 43, and an application server control unit 45.

The application server communication unit 41 has an interface for communicating with the user terminal 10, the data collection server 20, and the analysis server 30 through a communication network (NW). In particular, the application server communication unit 41 may receive information about products (eg, a list of products purchased together) according to the results of analyzing the behavior log from the analysis server 30.

The application server storage unit 43 stores data or programs executed or processed by the application server control unit 45. Additionally, the application server storage unit 43 may store information about products received from the analysis server 30 (for example, a list of products purchased together) temporarily or for a predetermined period of time.

The application server control unit 45 serves to overall control the operation of the application server 40, including the application server communication unit 41 and the application server storage unit 43, and may include a processor for this purpose. This application server control unit 45 stores information about products received from the application server 40 (for example, a list of products purchased together) in the collection server storage unit 43. And, if necessary, the collection server control unit 25 transmits information about products (for example, a list of products purchased together) to the user terminal 10 through the collection server communication unit 41.

Next, a method for providing information about a product according to an embodiment of the present invention will be described. Figure 6 is a flowchart illustrating a method for providing information about a product according to an embodiment of the present invention.

Referring to FIG. 6, the data collection server 20 may collect a plurality of behavior logs from a plurality of user terminals 10 in step S110. Here, the behavior log includes a user identifier (UID) and a product (item).

Next, the data collection server 20 transfers the plurality of behavior logs collected in step S120 to the analysis server 30. Then, the analysis server 30 analyzes the plurality of behavior logs received in step S130 through similarity calculation to generate information about the product. At this time, when calculating similarity, the data collection server 20 erases some of the plurality of behavior logs and performs similarity calculation through the remaining behavior logs. Accordingly, the speed of similarity calculation is improved.

Now, the similarity calculation method according to an embodiment of the present invention will be described in more detail.

The present invention calculates similarity using Jaccard similarity as shown in Equation 1 below.

That is, as shown in Equation 1, the value is calculated using the union of the action logs for the item, that is, product A and product B, as the denominator, and the intersection of the action logs for product A and product B as the denominator. Here, the union of the action logs for product A and product B is the action log in which the purchased product includes at least one of product A and product B, and the intersection of the action logs for product A and product B is the action log in which the purchased product includes at least one of product A and product B. It is a behavior log that includes both and product B. And the larger the calculated value, the higher the similarity. It is judged that the similarity between product A and product B is high.

For example, assume that data containing 6 action logs (L1 to L6) are as shown in Table 1 below.

behavior log User identifier Product A Product B L1 UID1 One 0 L2 UID2 One 0 L3 UID3 One 0 L4 UID4 0 One L5 UID5 0 One L6 UID6 0 One

As shown in the behavior logs L1, L2, and L3 in Table 1, if 3 users purchased only product A, and as shown in the behavior logs L4, L5, and L6, if 3 users purchased only product B, the similarity between product A and product B is mathematically According to Equation 1, (0/3+3-0) = 0. In this way, the behavior log of a user who purchased only one product among the behavior logs may not affect the similarity calculation results. Therefore, the present invention seeks to reduce unnecessary calculations on the relevant data (behavior log) during calculation by erasing behavior logs that have no effect on the similarity calculation result.

Meanwhile, assume that data including 200,000 behavior logs (L1 to L200,000) are as shown in Table 2 below.

behavior log User identifier Product A Product B L1 UID1 One 0 L2 UID2 One 0 L3 UID3 One 0 ?? ?? ?? ?? L100,000 UID100,000 One 0 L100,001 UID100,001 0 One L100,002 UID100,002 0 One L100,003 UID100,003 0 One ?? ?? ?? ?? L199,999 UID199,999 0 One L200,000 UID200,000 One One

For Table 2, 100,000 users purchased product A, as shown in action logs L1 to L100,000, and 199,999 users only purchased product B, as shown in action logs L100,001 to L199,999, and action log L200,000 As shown, assume that one user purchased both product A and product B.

In this case, the similarity between product A and product B results in 1/(100,001+100,000-1) = 1/200,000 according to Equation 1.

Here, if one user deletes all action logs of purchasing only one product, the similarity between product A and product B becomes 1/(1+1-1)=1 according to Equation 1. This causes distortion of the results.

In this way, if one of the plurality of behavior logs deletes all behavior logs in which one user purchased only one product, the results of the similarity calculation may be distorted.

As shown in action logs L1 to L6 in Table 1, if only two products A and B are considered, it can be determined that the user of each action log purchased only one product. However, since a user may have purchased products H, I, etc., if the user's behavior log is deleted, the similarity calculation results may be distorted.

Therefore, the present invention can specify the behavior log to be erased, that is, the object of erasure, according to the number of recommended products (RG) and the number of purchased products in the behavior log so as to reduce the amount of calculation without distorting the results of the similarity calculation. .

According to the method described above, the analysis server 30 calculates information about the product through similarity calculation.

Next, the analysis server 30 transmits the information about the product created in step S140 to the application server 40. Information about these products may be, for example, a list of products purchased together. The application server 40, which has received information about the product, can provide information about the product to the user terminal 10, if necessary.

Then, we will explain in more detail how to perform the similarity calculation by specifying an erase target so as to reduce the amount of calculation without distorting the result of the similarity calculation in step S130 described above. Figure 7 is a flowchart illustrating a high-speed similarity calculation method according to an embodiment of the present invention. The embodiment referring to FIG. 7 is for explaining step S130 of FIG. 6 in more detail.

Referring to FIG. 7, the preprocessing module 35a reads a plurality of behavior logs stored in the analysis server storage unit 32 in step S131, and converts the plurality of behavior logs into a user identifier (UID) and a product (item) in step S132. You can sort according to . For example, the preprocessing module 35a can create a table with user identifiers (UID) as columns and products (items) as rows.

Next, the preprocessing module 35a performs preprocessing to erase some of the plurality of behavior logs in step S133. At this time, the preprocessing module 35a determines the erasure ratio, which is the ratio of the behavior log to be erased from the total behavior log, depending on the amount of information about the product to be provided. Additionally, the preprocessing module 35a determines the behavior log (the number of purchased products in the behavior log) to be erased according to the determined erasure ratio. For example, when a user searches for a target product (TG), temporarily stores it (adds it to the shopping cart), or purchases it through the user terminal 10, the recommended product (RG), that is, the 'together purchased product', is selected. Let's assume a situation where you want to provide 'product' as information about the product. At this time, as the number of recommended products (RG), which is the amount of information about the product to be provided, increases, the deletion rate, which is the percentage of action logs to be erased, decreases. As the number of recommended products (RG) decreases, the purge rate, which is the percentage of action logs purged, increases. In addition, the preprocessing module 35a erases the behavior logs with one purchased product if the ratio of the total behavior logs with one purchased product is less than the erasure ratio, and if it is greater than the elimination ratio, the similarity calculation result is Because it affects the behavior, it is not possible to clear the action log with the number of purchased products being 1. Therefore, the behavior log to be erased is determined according to the determined erasure rate. Specifically, the pre-processing module 35a may erase the corresponding behavior log when the number of behavior logs indicating that one product has been purchased is less than the determined erasing ratio compared to the total behavior logs.

Next, the preprocessing module 35a sorts the plurality of behavior logs according to the product name in step S134. Then, the preprocessing module 35a groups a plurality of action logs according to the name of the product in step S135.

Next, the similarity calculation module 35b performs similarity calculation according to the name of the product in step S136. Here, the similarity calculation module 35b uses as the denominator the union of action logs including at least one of one product and another product, and as the numerator the intersection of action logs including both one product and the other product. Calculate the similarity between the product and other products. At this time, the operation speed is improved because the action log according to the purchase history of purchasing one product in advance has been erased.

Next, the information providing module 35c generates information about the product according to the similarity calculation result of the similarity calculation module 35b in step S137. Here, the information providing module 35c can generate information about a predetermined number of products in descending order of similarity according to the similarity calculated by the similarity calculation module 35b.

Information about these products may be, for example, a list of products purchased together. The information providing module 35c can transmit information about the generated product to the application server 40 through the analysis server communication unit 31.

Meanwhile, a method of providing information about products based on high-speed similarity calculation according to a more specific embodiment of the present invention will be described. Figure 8 is a flowchart illustrating a method of providing information about a product according to another embodiment of the present invention. Figure 9 is an example screen for explaining a method of providing information about a product according to another embodiment of the present invention. To emphasize, the embodiment referred to in FIGS. 8 and 9 is the step S130 of FIG. 6 and the more specific embodiment of FIG. 7 .

Referring to Figures 8 and 9, it is assumed that the analysis server control unit 35 receives a product recommendation request message from the application server 40 through the analysis server communication unit 31 in step S210. As shown in FIG. 9, the product recommendation request message is a message requesting a recommended product (RG) for the target product (TG). Accordingly, the product recommendation request message includes the number of target products (TG) and recommended products (RG). For example, when a user searches for, temporarily stores (adds to shopping cart), or purchases a target product (TG) through the user terminal (10) through a website, the application server (40) provides a recommended product (RG). ), that is, assume a situation where you want to show 6 'products purchased together', and in this case, the number of recommended products (RG) is 6.

Upon receiving the product recommendation request message, the pre-processing module 35a sorts the plurality of behavior logs stored in the analysis server storage unit 32, that is, the entire behavior log, by user in step S220. For example, as shown in Table 3 below, the preprocessing module 35a creates a table with the user's user identifier (UID) as the column (row) and the purchased product (item) as the row for the entire behavior log. You can.

behavior log User identifier Product A Product B Product C Product D ?? L1 UID1 One 0 One One ?? L2 UID2 0 One 0 One ?? L3 UID3 One 0 0 0 ?? L4 UID4 0 One One 0 ?? L5 UID5 0 0 0 One ?? L6 UID6 One One 0 One ?? ?? ?? ?? ?? ?? ??

Next, the preprocessing module 35a determines the erasure ratio, which is the ratio of behavior logs that can be erased from the entire behavior log, according to the number of recommended products (RG) in the product recommendation request message in step S230. Here, the extinction rate decreases as the number of recommended products increases. Intuitively, the extinction rate may be inversely proportional to the number of recommended products. For example, when the number of recommended products (RG) is 6 and the extinction rate is 10%, if the number of recommended products (RG) is 30, the elimination rate may be 5%. Next, the preprocessing module 35a determines the erase target according to the erase rate and the number of purchased products in the behavior log in step S240. For example, assume that the number of recommended products (RG) is 6 and the resulting extinction rate is 10%. The preprocessing module 35a calculates the ratio of behavior logs in which the number of products purchased by one user is two or less, as shown in L2, L3, L4, and L5 in Table 3 among all behavior logs. At this time, if the calculated ratio is 15%, the preprocessing module 35a excludes the behavior logs with two or less purchased products from being erased because the calculated ratio is greater than the erasure rate (10% < 15%). And, as shown in L3 and L5 in Table 3, the preprocessing module 35a calculates the ratio of the log data in which the number of products purchased by one user is one behavior log. At this time, if the calculated ratio is 8%, the preprocessing module 35a determines the behavior log with the number of purchased products as 1 to be erased because the calculated ratio is less than the erasure rate (10% > 8%).

Subsequently, the preprocessing module 35a erases the previously determined behavior log to be erased from the entire behavior log in step S250.

Next, the preprocessing module 35a sorts the entire behavior log from which the deletion target has been removed in step S260 according to the product name (item), groups it according to the product name (item), and then extracts the group for the target product (TG). do. For example, if the target product (TG) is product A, that is, in order to recommend 'products purchased together' with product A, all action logs including purchased product A are extracted. That is, in the case of an action log where the number of purchased products subject to erasure is one, data L3 is erased, so L1 and L6 can become a group for product A.

Next, the similarity calculation module 35b calculates the similarity between the target product (TG) and a plurality of other products according to Equation 1 in step S270. That is, the similarity calculation module 35b uses the union of action logs including both the target product (TG) and at least one other product as the denominator, and the intersection of action logs including both the target product (TG) and other products as the numerator. Calculate the similarity between the target product (TG) and other products.

For example, when the target product (TG) is product A, the similarity calculation module 35b calculates the similarity between product A and each of the other products (products B, C, D, E, ??).

The similarity calculation module 35b uses the union of action logs for the target product (TG), for example, product A and other products (products B, C, D, E, ??) as the denominator and the target product according to Equation 1. Similarity can be calculated using the intersection of (TG) and behavior logs for other products as the denominator. Here, the union of the action logs for the target product (TG) and other products is the action log in which the purchased product includes at least one of the target product (TG) and other products, and the behavior for the target product (TG) and other products The intersection of logs is a behavior log where the purchased product includes both the target product (TG) and other products.

Subsequently, the information providing module 35c derives the number (e.g., 6) of recommended products RG as recommended products in order of high similarity according to the similarity calculated by the similarity calculation module 55b in step S280. . Then, the information providing module 35c transmits the derived recommended product as information about the product to the application server 40 through the analysis server communication unit 31. Accordingly, the application server 40 can configure a web page screen as shown in FIG. 9 and provide it to the user terminal 10.

According to the present invention as described above, the speed of the similarity calculation is improved by specifying a plurality of behavior logs that do not affect the result of the similarity calculation as objects of deletion and erasing them in the pre-processing process, thereby reducing the load on the system. You can.

As described above, although this specification contains details of numerous specific implementations, these should not be construed as limitations on the scope of any invention or what may be claimed, but rather may be specific to particular embodiments of a particular invention. It should be understood as a description of features. Certain features described herein in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination. Furthermore, although features may be described as operating in a particular combination and initially claimed as such, one or more features from a claimed combination may in some cases be excluded from that combination, and the claimed combination may be a sub-combination. It can be changed to a variant of a sub-combination.

Likewise, although operations are depicted in the drawings in a particular order, this should not be construed as requiring that those operations be performed in the specific order or sequential order shown or that all of the depicted operations must be performed to obtain desirable results. In certain cases, multitasking and parallel processing may be advantageous. Additionally, the separation of various system components in the above-described embodiments should not be construed as requiring such separation in all embodiments, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products. You must understand that it is possible.

Specific embodiments of the subject matter described herein have been described. Other embodiments are within the scope of the following claims. For example, the operations recited in the claims can be performed in a different order and still achieve desirable results. As an example, the process depicted in the accompanying drawings does not necessarily require the specific depicted order or sequential order to achieve desirable results. In certain implementations, multitasking and parallel processing may be advantageous.

The present description sets forth the best mode of the invention and provides examples to illustrate the invention and to enable any person skilled in the art to make or use the invention. The specification prepared in this way does not limit the present invention to the specific terms presented. Accordingly, although the present invention has been described in detail with reference to the above-described examples, those skilled in the art may make modifications, changes, and variations to the examples without departing from the scope of the present invention.

Therefore, the scope of the present invention should not be determined by the described embodiments, but by the scope of the patent claims.

The present invention relates to a method for recommending products based on high-speed similarity calculation and a recommendation server for the same. Using the present invention, the similarity calculation is performed by erasing unnecessary action logs when calculating similarity used to derive recommended products. Speed can be improved. Accordingly, the system load can be reduced and the system speed can be improved. This can contribute to the development of the service industry, and in addition, the present invention has industrial applicability because it not only has sufficient potential for commercialization or sales, but also can be clearly implemented in reality.

10: User terminal 11: Communication department
13: input unit 15: display unit
17: storage unit 19: control unit
20: Data collection server 21: Collection server communication department
23: collection server storage unit 25: collection server control unit
30: Analysis server 31: Analysis server communication department
33: Analysis server storage unit 35: Analysis server control unit
35a: Preprocessing module 35b: Similarity calculation module
35c: Information provision module 40: Application server
41: Application server communication unit 43: Application server storage unit
45: Application server control unit

Claims (9)

A data collection server collecting a plurality of behavior logs, each including a user identifier and a product, from the user terminal and transmitting them to the analysis server;
The analysis server deleting some of the plurality of behavior logs;
Calculating similarity by the analysis server performing a similarity calculation on a plurality of partially erased behavior logs;
generating, by the analysis server, information about a product according to the similarity calculation result; and
Including a step of the analysis server transmitting information about the product to the application server,
The step of erasing some of the plurality of action logs is
A method for providing information about a product, comprising determining an erasure rate according to the amount of information about the product to be provided, and determining an erasure target according to the determined erasure rate and the number of purchased products in the behavior log. delete delete ◈Claim 4 was abandoned upon payment of the setup registration fee.◈ According to paragraph 1,
The step of calculating the similarity by performing the similarity calculation is
The denominator is the union of action logs that includes one product and at least one of the other products,
With the intersection of the action log containing both one product and the other product as the numerator,
A method for providing information about a product, characterized in that calculating the similarity between one product and the other product. ◈Claim 5 was abandoned upon payment of the setup registration fee.◈ According to paragraph 4,
The step of generating information about the product according to the similarity calculation results is
A method for providing information about a product, characterized in that, according to the calculated similarity, a predetermined number of products are derived as information about the product in the order of the highest similarity. Analysis server communication unit for communication with the application server;
When a recommended product request message including the target product and the number of recommended products is received from the application server through the analysis server communication unit, a user identifier and a plurality of behavior logs including purchased products corresponding to the user identifier are selected for deletion. a preprocessing module that specifies and erases the specified erase target from a plurality of behavior logs; and
It includes a similarity calculation module that calculates the similarity with the target product using a plurality of behavior logs from which the deletion target has been erased,
The preprocessing module is
A device for providing information about products, wherein a deletion rate is calculated according to the number of recommended products, and a deletion target is determined according to the calculated deletion rate and the number of purchased products in the behavior log. delete delete delete

Images