KR20190100789A - Method for recommending information on websites customized to clients' log by using personalization algorithm and server using the same - Google Patents

Abstract

The present invention relates to a method of recommending a customer-log-customized website through a personalization algorithm. According to the present invention, a personalization algorithm includes the following steps of: forming a website tag table by generating a tag value and a purchase tag by website from a purchase history of a customer in regard to information about every website purchased by each customer; calculating similarity between websites based on the website tag table; forming a customer log tag table based on log information obtained from customers; selecting a predetermined number of high-ranking websites by customer from the customer log tag table; and suggesting a high-ranking website having high similarity to the selected websites as a personalized consumption-customized website based on the calculated similarity and a predetermined number of the selected websites.

Description

How to recommend customized site for customer log using personalization algorithm and server for personalization algorithm

The present invention relates to a method for recommending a customer log customized site using a personalization algorithm, and more specifically, (a) a personalization algorithm server, based on purchase information that customers purchase using an arbitrary site, respectively. Generating a site tag table listing site-specific site purchase tags for all site information purchased by each customer, and tag values formed based on the number of purchases of each specific customer on the specific site, for a customer of; (b) the personalization algorithm server calculates the similarity between the two randomly selected sites using tag values of all customers for each of two randomly selected sites among all the sites in the site tag table. Generating a site similarity table indicating the similarity value between the respective sites; (c) for each customer and each customer based on log information including at least some of the likelihood display, bookmarking and click actions performed by the personalization algorithm server on at least one particular site acquired from the customer. Tag values formed based on the number of logs including the number of processing of at least a portion of each of the log information tags formed by the specific sites and the likelihood display, bookmark processing, and click actions corresponding to the log information tags. Generating a customer log tag table; (d) selecting, by the personalization algorithm server, sites belonging to an upper predetermined rank for each customer based on tag values for each site in the customer log tag table; And (e) the personalization algorithm server associated with each customer for each customer, and the sites having a high similarity value with the predetermined rank higher rank sites and the higher rank sites selected in step (d). The present invention relates to a method and a personalization algorithm server including a; presenting to the customer log customized recommendation site optimized for the customer.

Due to the development of the Internet, e-commerce using the Internet is exploding. As the amount of information provided through the Internet increases exponentially, users have to spend a lot of time and effort to search, search, and compare information. In other words, the era of information glut has come and consumers have to spend more time making choices and decisions because of the excess of information, services and production.

In particular, when using an online shopping mall, a user searches directly by inputting an item type or a site name that he / she wants to purchase through an internet search, and then browses the list presented by the search site or shopping mall one by one or follows some links presented. After looking at, proceed with the purchase.

On the other hand, as user requirements become more demanding and personalized requirements increase, services that provide user-specific information based on data analyzing individual patterns have been proposed in various ways.

As one of the existing methods, the method described in "Customized shopping information system and method through analysis of user tendency" disclosed in Korean Patent Publication No. 10-1419504 may provide effective shopping information through analysis of purchase tendency of each user. In order to provide customized shopping information through user disposition analysis. In the above conventional technology, the ontology module generates the at least one product tag and at least one propensity tag for each product tag by ontology processing the collected information, and the product tag generated through the ontology module. Various purchasing tendencies of customers are analyzed using tag matching storage module which maps with tags and stores them in information database.

However, the existing analysis method provides customized shopping information based on the analysis of each user's propensity and specific terminology, so that only the individual customer's inclination is provided and the term-based matching method is optimized for the customer. There is a limit to providing.

The present invention aims to solve all the above-mentioned problems.

In order to provide an optimized site for a specific customer, an object of the present invention is to provide an optimized site with high reliability to a specific customer based on the log information of the customer related to the site obtained from the customer even for those customers who are not in the purchase history. do.

The present invention also aims to provide information optimized for each customer by integrating information of multiple or all customers obtained from multiple customers.

The characteristic structure of this invention for achieving the objective of this invention mentioned above, and realizing the characteristic effect of this invention mentioned later is as follows.

According to an embodiment of the present invention, a method of recommending a customer log personalized site using a personalization algorithm includes: (a) a personalization algorithm server based on purchase information purchased by a customer using an arbitrary site; Generating, for a customer, a site tag table listing site-specific site purchase tags for all site information purchased by each customer and tag values formed based on the number of purchases of each particular customer on the specific site; (b) the personalization algorithm server calculates the similarity between the two randomly selected sites using tag values of all customers for each of two randomly selected sites among all the sites in the site tag table. Generating a site similarity table indicating the similarity value between the respective sites; (c) for each customer and each customer based on log information including at least some of the likelihood display, bookmarking and click actions performed by the personalization algorithm server on at least one particular site acquired from the customer. Tag values formed based on the number of logs including the number of processing of at least a portion of each of the log information tags formed by the specific sites and the likelihood display, bookmark processing, and click actions corresponding to the log information tags. Generating a customer log tag table; (d) selecting, by the personalization algorithm server, sites belonging to an upper predetermined rank for each customer based on tag values for each site in the customer log tag table; And (e) the personalization algorithm server associated with each customer for each customer, and the sites having a high similarity value with the predetermined rank higher rank sites and the higher rank sites selected in step (d). And presenting them as a customer log customized recommendation site optimized for the customer.

In an embodiment, the step (d) may include tag values of all log information tags corresponding to the same site for each customer, for one or more log information tags corresponding to the same site for each customer. A sum tag value, which is added with reference to the type of information tag, is generated, and sites belonging to the upper predetermined rank for each customer are selected based on the sum tag value.

In one embodiment, step (d) multiplies corresponding tag values by predetermined weights different from each other according to the log information tag for each site, and adds them.

In one embodiment, the similarity value calculation between the sites in step (b) is

Is calculated through, where

a = (a ₁ , a ₂ ,…, a _m ): tag values of a site in the site tag table

b = (b ₁ , b ₂ ,…, b _m ): Tag values of other sites in the site tag table

m is the total number of customers with site purchases.

In an embodiment, the step (e) may include the information of the higher rank sites selected through the step (d) for each of the sites and the sites having a similarity or higher than the predetermined value obtained in the step (b). Based on the number of similar sites (Voting value) for each of the selected top ranking sites and the similarity value of the similarity with the similar sites, using the largest similarity value, the score (Score) is calculated And, the predetermined number of sites in the order of the recommendation value is presented to the customer log customized recommendation site.

In one embodiment, the step (e) may comprise a first site, which is any one of the higher rank sites selected through the step (d), and one or more agents that are similar to the first site. Based on the similarities between the two sites, a total value of the second sites is calculated, a maximum value of the similarity values between the first site and the second sites is calculated, and then the second value is calculated. For each of the sites, a recommendation value for the customer is calculated.

In one embodiment, the recommended value is

Where cosine is the largest of similarity values with related sites of a particular site, voting_weight is a weighting constant, and voting is a second, similar to the first site, wherein the similarity value is greater than zero. The number of sites, max (voting, for each customer) is the maximum of each customer's voting values.

According to another feature of the invention, the personalization algorithm server for recommending a customer log customized site, the communication unit for receiving the purchase history of each customer and / or log information about the site to the customer; And (i) site-specific site purchase tags for each customer, all site information purchased by each customer, based on purchase information purchased by the customer using any site, and for each particular customer, the specific site. Generating a site tag table listing the tag values formed based on the number of purchases in the; (ii) using the tag values of all customers for each of the two sites randomly selected from all the sites in the site tag table, calculating the similarity between the two randomly selected sites to calculate the similarity between the respective sites. Creating a site similarity table representing the values; (iii) based on log information including at least some of the likelihood display, bookmarking, and click actions performed on at least one particular site acquired from the customer, categorized by each customer and by the particular site for each customer And generate a customer log tag table listing tag values formed on the basis of each log information tag formed and the number of logs including the number of processing of at least some of a crush display, bookmark processing, and a click action corresponding to the log information tag. Process; (iv) selecting a site belonging to an upper predetermined rank for each customer based on tag values for each site in the customer log tag table; And (v) optimizing the top sites of the predetermined ranks related to each customer for each of the customers, and the sites having high similarity values to the selected top rank sites in the process. And a processor for performing the process of presenting the customer log customized recommendation site.

The present invention provides a personalization algorithm that provides an optimized site with high reliability, based on log information, including behavioral information such as visits to customers, even for customers who do not have a purchase history when presenting a site optimized for customers. do.

In addition, the present invention provides a personalization algorithm that collects from all customers without using only the tags (attributes) of the customer, and provides an optimal site for the customer in the purchasing characteristics of the customers with similar characteristics.

1 is a diagram illustrating the basic concept of a personalization algorithm according to the present invention.
2 is a flowchart of a method for providing an optimized site using a personalization algorithm using log information according to the present invention.
3 shows how the similarity between sites is calculated in the site tag table.
4 shows an example of obtaining log information for each site from member information.
5 is a flowchart illustrating a process of generating a tag value of a customer log tag table according to the present invention.
6 is a schematic structural diagram of a personalization algorithm server according to the present invention.

DETAILED DESCRIPTION The following detailed description of the invention refers to the accompanying drawings that show, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different but need not be mutually exclusive. For example, certain shapes, structures, and characteristics described herein may be embodied in other embodiments without departing from the spirit and scope of the invention with respect to one embodiment. In addition, it is to be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. The following detailed description, therefore, is not to be taken in a limiting sense, and the scope of the present invention, if properly described, is defined only by the appended claims, along with the full range of equivalents to which such claims are entitled. Like reference numerals in the drawings refer to the same or similar functions throughout the several aspects.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement the present invention.

1 is a diagram illustrating the basic concept of a personalization algorithm according to the present invention.

Referring to FIG. 1, the personalization algorithm according to the present invention uses tag values based on log data such as the number of clicks, the number of bookmarks performed, the number of liked clicks, etc. that a customer has performed on a specific site based on the activity-based data of the customers. Form the reflected customer log tag table. In addition, the site tag table is generated by generating purchase tag and tag value for each site based on the information on all sites purchased by each customer from the purchase history of all customers. Calculate the similarity between.

Then, the top predetermined number of sites are selected for each customer from the customer log tag table, and sites similar to the selected sites are selected to recommend sites optimized for each customer.

Here, the purchase data of the customers may be acquired through card usage history.

2 is a flowchart of a method for providing an optimized site using a personalization algorithm using log information according to the present invention.

Referring to FIG. 2, the personalization algorithm using the customer log tag according to the present invention forms a site tag table by generating site-specific purchase tags and tag values for all site information purchased by each customer from the purchase history of the customer. Step (S10), calculating a similarity between each site using the site tag table (S20), forming a customer log tag table based on the log information obtained from customers (S30), customer log tag table Selecting a predetermined number of sites per customer from the customer (S40), and based on the calculated similarity and the predetermined number of selected sites, the high ranking similarity with sites selected for each customer personalized consumption personalization Presenting to the site (S50).

FIG. 3 shows how similarity between sites is calculated in the site tag table, and FIG. 4 shows an example of obtaining log information for each site from member information.

5 is a flowchart illustrating a process of generating a tag value of a customer log tag table according to the present invention.

Referring to FIG. 5, in step S30 of FIG. 2, a site key extraction step S31 corresponding to log information for each customer from customer information and log information of each customer, tag values set differently according to log information for log information And a step (S33) of forming a customer log tag table by summing corresponding tag values for each customer and for each specific site for each customer.

6 is a schematic diagram of a personalization algorithm server according to the present invention.

Referring to FIG. 6, the personalization algorithm server 100 according to the present invention includes a communication unit 10 and a processor 20.

Hereinafter, a method for recommending a site using a personalization algorithm according to the present invention will be described in detail with reference to FIGS. 1 to 6.

First, the communication unit 10 performs a function for receiving the purchase details and log activity details of each customer, and the processor 20 functions to perform a process according to each step of FIGS. 2 and 5.

First, through the communication unit 10 of the personalization algorithm server 100, the customer's purchase history and customer information for personalization and the log information history including the log activity history and the customer's log activity history, the processor 20 the customer In step S10, a site tag table is generated by generating site-specific purchase tags and tag values for all site information purchased by each customer from the purchase history of the customer. For each customer, based on the purchase information purchased using the site, a site-specific site purchase tag (e.g., site key) for all site information purchased by each customer and the number of purchases from that particular site by each particular customer. A site tag that lists tag values (for example, the number of purchases made on a site corresponding to a specific site key) To form a label.

For example, the processor 20 receives a plurality of customer-specific purchase history information for all customers (for example, n customers) through the communication unit 10, and each purchase history is a purchase history of a certain customer. Tag the purchase personalization tag (that is, the site purchase tag) for each purchase and whether it is a purchase related to a specific site from each purchase history, and add up the number of purchases associated with the specific site purchase tag for each customer and set the tag value. To form a site tag table by associating site purchase tags with tag values for each customer.

Table 1 shows examples of site purchase tags and their tag values obtained through some purchase history for customers 1 to 5 for site keys "FS1600001" and "FS160002".

Customer number Site purchase tags CNT
(count) Customer1 Site Purchase_FS1600001_Number 4 Customer2 Site Purchase_FS1600001_Number 3 Customer3 Site Purchase_FS1600001_Number 2 Customer4 Site Purchase_FS1600001_Number 6 Customer1 Site Purchase_FS1600002_Number 3 Customer2 Site Purchase_FS1600002_Number One Customer4 Site Purchase_FS1600002_Number 4 Customer5 Site Purchase_FS1600002_Number 8

Referring to Table 1, the site tag table is formed by using the purchase information of customers 1 to 5, and the purchase history of the specific number of purchases of the customer 1 (for example, the purchase number 01) is calculated. If it is confirmed that the purchase history for the site that is "FS1600001" is set, the count of customer 1's site purchase tag "Site Purchase_FS1600001_Number" is set to 1, and if the specific number of purchases of Customer 1 (for example, purchase number 02), If the purchase history of the user is again confirmed as the purchase history for the site whose site key is "FS160001", the count of the site purchase tag "site purchase_FS1600001_" of the customer 1 is changed to 2. In this way, if customer 1's entire purchase history is reviewed and the number of purchases for the site with site key "FS1600001" is four times, then the tag for customer 1's site purchase tag "Site Purchase_FS1600001_number" The value is set to 4. Similarly, if customer 1's entire purchase history is reviewed and the number of purchases for the site with site key "FS1600002" is three times, then the tag value for customer 1's site purchase tag "Site Purchase_FS1600002_Number" Is set to 3.

Similarly, if customer 2's entire purchase history is reviewed and the number of purchases for the site with site key "FS1600001" is three times, then the tag value for customer 2's site purchase tag "Site Purchase_FS1600001_number" Is set to 3. If the number of purchases for the site having the site key "FS1600002" is one time, the tag value for the site purchase tag "Site Purchase_FS1600002_Number" of the customer 2 is determined. Is set to one.

In this way, for all purchased customers, the entire purchase history is reviewed for each customer, the number of times per site key is analyzed, the site purchase tag and the number of purchases corresponding to the site purchase tag are set as the tag value. Form a table.

Then, the processor 20 performs a process of calculating similarity between the sites using the site tag table of Table 1 (S20).

Specifically, the processor 20 calculates the similarity between the two randomly selected sites, using the tag values of all customers for each of two randomly selected sites among all the sites in the site tag table of Table 1. To generate a site similarity table representing the similarity values between the sites.

3 shows how the similarity between sites is calculated in the site tag table.

Referring to FIG. 3, the similarity between sites is calculated through the site tag table of Table 1 and the same table. For example, the site purchase tag " site purchase_ " in one site tag table (first site tag table) is calculated. The site purchase tag "Site Purchase_FS1600002" in the site tag table that differs from the customer's tag values for FS1600001_number "(where the second site tag table is the same table as the first site tag table). Using the tag values of customers for _count ", the similarity of the site for site key" FS1600001 "and the site for site key" FS1600002 "is calculated.

In this way, it is possible to obtain site similarity between each other for each site. If the site in the site s dog tag table exists, s ² could be to calculate the similarity of values.

Meanwhile, the processor 20 calculates similarity values between any two sites by using tag values of the site tag table of Table 1 using the cosine similarity function of Equation 1 below.

Where cos (θ) represents the cosine similarity function,

a = (a ₁ , a ₂ ,..., a _m ) represents tag values of any one site (first site) in the site tag table. That is, if there are m customers for any one site, for each customer there are tag values for the any one site and a represents the tag values. And b = (b ₁ , b ₂ ,..., B _m ) is a tag value of a tag table of another site (second site) in the site tag table, and m represents the total number of customers who have a site purchase.

By using the site tag table of Table 1 to obtain the similarity values between each through Equation 1, it can be calculated as shown in the similarity result table (Table 2) below.

First site key Second site key Similarity
(cosine) FS1600002 FS1600002 One FS1600002 FS1603708 0.0483368 FS1600004 FS1600004 One FS1600004 FS1602529 0.2041241 FS1600004 FS1601505 0.1581139 FS1600004 FS1603708 0.0206725 FS1600004 FS1603904 0.0117671 FS1603708 FS1603708 One FS1603708 FS1600002 0.0483368 FS1603708 FS1601048 0.0483368 FS1603708 FS1603978 0.0483368 FS1603708 FS1604053 0.0483368 FS1603708 FS1603990 0.0432338 FS1603708 FS1600284 0.0402671 FS1603708 FS1603926 0.0399273 FS1603708 FS1600794 0.0354544 FS1603708 FS1601261 0.0346641 FS1603708 FS1603271 0.0341793 FS1603708 FS1601105 0.0319787 FS1603708 FS1601432 0.0306303 …

Referring to Table 2, by way of example, the similarity between the site keys "FS1600002" and "FS1600002" is the same site, so the value is 1, and the similarity between the site keys "FS1600002" and "FS1600008" is 0.0483368. , The similarity between FS1600004 "and" FS1600004 "is the same site, so a value of 1 is obtained, and the similarity between the site keys" FS1600004 "and" FS1602529 "is 0.2041241.

Referring back to FIG. 2, the processor 20 performs a step S30 of forming a customer log tag table based on log information obtained from customers. Specifically, the processor 20 is based on the log information including at least some of the likelihood display, bookmark processing and click actions performed on at least one specific site acquired from the customer, for each of the customer and for each customer A customer who lists tag values formed based on each log information tag formed by the specific site and the number of logs including the number of processing of at least a part of a crush display, bookmark processing, and a click action corresponding to the log information tag. Create a log tag table.

This process is illustrated in more detail in FIG.

Referring to the process of forming the customer log tag table of Figure 5, the process of creating a customer log tag table, first extracting the site key corresponding to the log information for each customer from the customer information and log information of each customer (S31) To start. Referring to Figure 4, from the customer information, the customer clicks 'like' for a specific site, the information 'visited' a specific site, the information 'marked' for a specific site is obtained, the 'like' Extracts the site key from the information that you clicked on, and also extracts the site key from which you visited, and the site key from which you've bookmarked the site. Check it by extracting it.

It then forms a customer log tag table by counting what log information (for example, likes, bookmarks, clicks, etc.) is present for every customer and for each site, and how many times there are corresponding actions.

An example of the generated customer log tag table is shown in Table 3 below.

Customer number Site key Log information tag CNT
(count) Customer1 FS1600002 good One Customer1 FS1600002 Bookmark 2 Customer1 FS1600002 click 5 Customer1 FS1603708 good One Customer1 FS1600004 good 2 Customer2 FS1600251 good 2 Customer2 FS1600380 Bookmark 3 Customer2 FS1600225 Bookmark One Customer2 FS1601700 Bookmark One Customer3 FS1600002 good One Customer3 FS1603708 click 6 Customer3 FS1600004 click 10 Customer3 FS1600380 good 4 …

Referring to Table 3, Customer1 clicks 'Like' once for the site corresponding to the site key "FS1600002", performs two "Bookmarks", "clicks" five visits to the site, Indicates that the user liked "like" once for the site corresponding to the site key "FS1603708". As such, the log tag table is generated by listing the log information tags corresponding to the log information and the tag values corresponding to the number of times for each customer for every customer.

Then, the processor 20 performs step S32 of assigning tag values differently set according to the log information for each log information. For example, the number of times of performing 'like' and 'bookmark' once may be set at the same rate as 'click', that is, 5 visits. Accordingly, the changed tag value is shown in Table 4 below.

Customer number Site key Log information tag CNT
(count) Tag value Customer1 FS1600002 good One One Customer1 FS1600002 Bookmark 2 2 Customer1 FS1600002 click 5 One Customer1 FS1603708 good One One Customer1 FS1600004 good 2 2 Customer2 FS1600251 good 2 2 Customer2 FS1600380 Bookmark 3 3 Customer2 FS1600225 Bookmark One One Customer2 FS1601700 Bookmark One One Customer3 FS1600002 good One One Customer3 FS1603708 click 6 One Customer3 FS1600004 click 10 2 Customer3 FS1600380 good 4 4 …

Referring to the bolded parts of Table 3 and Table 4, the tag value of the log information tag "click" for the customer 1's site key "FS1600002" is set to 1 based on the CNT (count value) 5, The tag value of the log information tag "click" for the site key "FS1603708" is set to 1 based on CNT (count value) 6, and the tag value of the log information tag "click" for the site key "FS1600004" of customer 3 Is set to 2 based on CNT (count value) 10.

Then, the processor 20 forms a customer log tag table by summing corresponding tag values for each customer and for each specific site for each customer (S33). Specifically, in Table 4, when there are three identical log information tags for the site key "FS1600002" for the customer 1, the tag values for the same site key "FS1600002" are summed. In this way, all the tag values for the same site for all customers are added together to form a customer log tag table by having one tag value for each site.

This exemplary result is shown in Table 5 below.

Customer number Site key Tag Values (Summary Tag Values by Site) Customer1 FS1600002 4 Customer1 FS1600004 2 Customer1 FS1603708 One Customer2 FS1600225 One Customer2 FS1600251 2 Customer2 FS1600380 3 Customer2 FS1601700 One Customer3 FS1600002 One Customer3 FS1600004 2 Customer3 FS1600380 4 Customer3 FS1603708 One …

Referring back to FIG. 2, the processor 20 performs a step S40 of selecting a predetermined number of sites for each customer from the customer log tag table. In detail, the processor 20 selects sites belonging to the upper predetermined ranking for each customer based on the tag values for each site in the customer log tag table. For example, the top three top four sites may be selected in the order of high tag values for each customer.

If the top three sites were selected for customer 1 in Table 5, and the site groups were selected as "FS1600002", "FS1600004", and "FS1603708", for customer 1, select only the similarity results for the sites in Table 2. To form a recommendation site list table for the customer 1. Exemplary results are shown in Table 6 below.

Customer number First site key Second site key Similarity
(cosine) Customer1 FS1600002 FS1600002 One Customer1 FS1600002 FS1603708 0.0483368 Customer1 FS1600004 FS1600004 One Customer1 FS1600004 FS1602529 0.2041241 Customer1 FS1600004 FS1601505 0.1581139 Customer1 FS1600004 FS1603708 0.0206725 Customer1 FS1600004 FS1603904 0.0117671 Customer1 FS1603708 FS1603708 One Customer1 FS1603708 FS1600002 0.0483368 Customer1 FS1603708 FS1601048 0.0483368 Customer1 FS1603708 FS1603978 0.0483368 Customer1 FS1603708 FS1604053 0.0483368 Customer1 FS1603708 FS1603990 0.0432338 Customer1 FS1603708 FS1600284 0.0402671 Customer1 FS1603708 FS1603926 0.0399273 Customer1 FS1603708 FS1600794 0.0354544 Customer1 FS1603708 FS1601261 0.0346641 Customer1 FS1603708 FS1603271 0.0341793 Customer1 FS1603708 FS1601105 0.0319787 Customer1 FS1603708 FS1601432 0.0306303 …

In Table 6, only the customer 1 is shown, but all of the selected top sites and similarity results of the selected top sites and other sites may be formed in the table.

Then, the processor 20 performs a step (S50) of presenting a high ranking site with a high degree of similarity with sites selected for each customer based on the calculated similarity and a predetermined number of selected sites (S50). .

Specifically, the customer log customized recommendation site optimized for each customer by the customer of the top sites of the predetermined ranks selected in step S40 and the high similarity value to the top sites of the selected predetermined ranks in relation to each customer for each customer In this case, the processor 20, based on the information of the predetermined rank higher sites selected through the step S40 for each site and the sites having a similarity or more than a predetermined value obtained in the step S20, Score (Score) by using the largest similarity value of (i) the number of similar sites for each of the selected top ranking sites and (ii) similarity values with the similar sites. Calculate and present a predetermined number of sites to the customer log customized recommendation site in order of the recommendation value being larger.

For example, as shown in Table 6, between the first site and any one or more second sites that are similar to the first site, any one of the higher rank sites selected through the step S40. After calculating the total number of the second sites (Voting value), and calculating the maximum value of the similarity value between the first site and the second sites, based on the similarity of Calculate a recommendation value for the customer.

Referring to the bolded part of Table 6, the counting number for the site key "FS1600002" among the second sites similar to the first sites selected in step S40 for the customer 1, that is, the site key " The total value of FS1600002 "is 2, and the counting number for the site key" FS1603708 "among the second sites similar to the first sites selected in step S40 for the customer 1, that is, at the second sites The total number of site keys " FS1603708 " is three. In this way, when the total number of each of the second sites for all customers including the customer 1 is calculated, a result for the total number of the second sites for each customer may be generated as shown in Table 7 below.

Customer number Second site key VOTING
(Total number) Maximum similarity
(cosine) Customer1 FS1603708 3 One Customer1 FS1600002 2 One Customer1 FS1600004 One One Customer1 FS1602529 One 0.204124145 Customer1 FS1601505 One 0.158113883 Customer1 FS1601048 One 0.048336824 Customer1 FS1603978 One 0.048336824 Customer1 FS1604053 One 0.048336824 Customer1 FS1603990 One 0.04323377 Customer1 FS1600284 One 0.040267136 Customer1 FS1603926 One 0.039927341 Customer1 FS1600794 One 0.035454426 Customer1 FS1601261 One 0.034664133 Customer1 FS1603271 One 0.034179296 … … … … Customer2 FS1603990 One 0.04323377 Customer2 FS1600284 One 0.040267136 …

Meanwhile, referring to Table 7, the processor 20 also obtains a similarity value between the site 1 and the site 2 when calculating the total number of the second site, and has the largest value among the similarity values associated with the second site. Come. That is, the similarity values associated with the site key "FS1603708" in Table 7 have the values "0.0483368", "0.0206725", and "1" in Table 6, which results in the largest value of "1", and in Table 7 The similarity value associated with site key "FS1600002" has the values "1" and "0.0483368" in Table 6, resulting in the largest value "1", and the similarity value associated with site key "FS1602529" in Table 7. In Table 6, only "0.2041241" has a value, and the value is taken as it is (as shown in Table 7, "0.204124145", but the rounded value is shown in Table 6).

Then, the processor 20 uses Equation 2 below based on the number of second sites in Table 7, that is, the number of sites similar to the sites selected in S40 and the maximum similarity value of each of these sites. To calculate the recommended value.

Where cosine is the largest of similarity values with related sites of a particular site, voting_weight is a weighting constant, and voting is the number of second sites similar to the first site, wherein the similarity value is greater than zero, max (voting, for each customer) is the maximum of each customer's voting values.

Using Table 7 to calculate the recommended value for the customer and specific site, it is shown in Table 8 below.

Customer number Second site key VOTING
(Total number) Maximum similarity
(cosine) Recommended value Customer1 FS1603708 3 One One Customer1 FS1600002 2 One 0.999917 Customer1 FS1600004 One One 0.999833 Customer1 FS1602529 One 0.204124145 0.672268 Customer1 FS1601505 One 0.158113883 0.630758 Customer1 FS1601048 One 0.048336824 0.469577 Customer1 FS1603978 One 0.048336824 0.469577 Customer1 FS1604053 One 0.048336824 0.469577 Customer1 FS1603990 One 0.04323377 0.456753 Customer1 FS1600284 One 0.040267136 0.44877 Customer1 FS1603926 One 0.039927341 0.447829 Customer1 FS1600794 One 0.035454426 0.434836 Customer1 FS1601261 One 0.034664133 0.432415 Customer1 FS1603271 One 0.034179296 0.430909 … … … … … Customer2 FS1603990 One 0.04323377 0.45849 Customer2 FS1600284 One 0.040267136 0.450601 …

The processor 20 presents a predetermined number of sites as the customer log customized recommendation site in the order of increasing the recommendation value based on the recommendation value calculated through Equation 2 above.

For example, a site that is presented to Customer1 as a customized site based on log information (if recommending the top four sites) would have the top site "FS1603708", " FS1600002 "," FS1600004 ", and" FS1602529 ".

The present invention, through the personalization algorithm as described above, when presenting the site optimized for the customer, even for customers who do not have a purchase history, optimized site with high reliability based on the log information including behavior information, such as customer visits to the site of the customer Provides a personalization algorithm to provide. In addition, the present invention can provide a personalization algorithm that collects from all customers without using only the tags (attributes) of the customer, and provides an optimal site for the customer in the purchasing characteristics of customers with similar characteristics.

Embodiments according to the present invention described above may be implemented in the form of program instructions that may be executed by various computer components, and may be recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the computer-readable recording medium may be those specially designed and configured for the present invention, or may be known and available to those skilled in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROMs, DVDs, and magneto-optical media such as floptical disks. media), and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform the process according to the invention, and vice versa.

Although the present invention has been described by specific embodiments such as specific components and the like, but the embodiments and the drawings are provided to assist in a more general understanding of the present invention, the present invention is not limited to the above embodiments. For those skilled in the art, various modifications and variations can be made from these descriptions.

Accordingly, the spirit of the present invention should not be limited to the above-described embodiments, and all of the equivalents or equivalents of the claims, as well as the appended claims, fall within the scope of the spirit of the present invention. I will say.

Claims (14)

In the method of recommending customer log customized site using personalization algorithm,
(a) The personalization algorithm server, for each customer, based on the purchase information that the customer purchased from any site, site-specific site purchase tags for each customer, and for each specific customer, Generating a site tag table listing tag values formed based on the number of purchases on the specific site;
(b) the personalization algorithm server calculates the similarity between the two randomly selected sites using tag values of all customers for each of two randomly selected sites among all the sites in the site tag table. Generating a site similarity table indicating the similarity value between the respective sites;
(c) for each customer and each customer based on log information including at least some of the likelihood display, bookmarking and click actions performed by the personalization algorithm server on at least one particular site acquired from the customer. Tag values formed based on the number of logs including the number of processing of at least a portion of each of the log information tags formed by the specific sites and the likelihood display, bookmark processing, and click actions corresponding to the log information tags. Generating a customer log tag table;
(d) selecting, by the personalization algorithm server, sites belonging to an upper predetermined rank for each customer based on tag values for each site in the customer log tag table; And
(e) the personalization algorithm server, for each of the customers, the sites having higher similarity values to the higher rank sites selected in the step (d) and the higher rank sites selected in the step (d). Presenting the customer log customized recommendation site optimized for the customer;
Site recommendation method comprising a. The method of claim 1,
In step (d), for one or more log information tags corresponding to the same site for each customer, the tag values of all log information tags corresponding to the same site for each customer are referred to log information tag types. Generating a sum tag value, and selecting sites belonging to the upper predetermined rank for each customer based on the sum tag value. The method of claim 2,
In the step (d), the site recommendation method, characterized in that by multiplying the corresponding tag value by a predetermined weight different from each other according to the log information tag for each site. The method of claim 1,
The similarity value calculation between sites in step (b) is

Is calculated through, where
a = (a ₁ , a ₂ ,…, a _m ): tag values of a site in the site tag table
b = (b ₁ , b ₂ ,…, b _m ): Tag values of other sites in the site tag table
m is the total number of customers who have purchased a site. The method of claim 4, wherein
The step (e) is based on the information of the predetermined rank higher sites selected through the step (d) for each site and the sites having a similarity or more than the predetermined value obtained in the step (b), Using the number of similar sites (Voting value) and the similarity value with the similar sites for each site of the selected top rank sites, a recommendation value (Score) is calculated, and the recommendation value Site recommendation method characterized in that the predetermined number of sites presented in the large order to the customer log customized recommendation site. The method of claim 5,
The step (e) is performed between a first site, which is any one of the higher rank sites selected through the step (d), and one or more second sites that are similar to the first site. Based on the similarities, the total number of the second sites is calculated, a maximum value of the similarity values between the first site and the second sites is calculated, and then, for each of the second sites. Site recommendation method characterized in that for calculating the recommendation value for the customer. The method of claim 6,
The recommended value is the following formula

Where cosine is the largest of similarity values with related sites of a particular site, voting_weight is a weighting constant, and voting is a second, similar to the first site, wherein the similarity value is greater than zero. The number of sites, max (voting, each customer) is a site recommendation method, characterized in that the maximum value of each customer's voting value. In the personalization algorithm server that recommends customer log customized site,
Communication unit for receiving each customer's purchase history and / or log information about the site to the customer; And
(i) site-specific site purchase tags for each customer, all site information purchased by each customer, and each specific customer, based on the purchase information purchased by the customer using any site; Generating a site tag table listing tag values formed based on the number of purchases of the; (ii) using the tag values of all customers for each of the two sites randomly selected from all the sites in the site tag table, calculating the similarity between the two randomly selected sites to calculate the similarity between the respective sites. Creating a site similarity table representing the values; (iii) based on log information including at least some of the likelihood display, bookmarking, and click actions performed on at least one particular site acquired from the customer, categorized by each customer and by the particular site for each customer And generate a customer log tag table listing tag values formed on the basis of each log information tag formed and the number of logs including the number of processing of at least some of a crush display, bookmark processing, and a click action corresponding to the log information tag. Process; (iv) selecting a site belonging to an upper predetermined rank for each customer based on tag values for each site in the customer log tag table; And (v) optimizing the top sites of the predetermined ranks related to each customer for each of the customers, and the sites having high similarity values to the selected top rank sites in the process. And a processor for performing a process of presenting a customer log customized recommendation site. The method of claim 8,
The (iv) process may refer to log information tag types for tag values of all log information tags corresponding to each same site for each customer, for one or more log information tags corresponding to the same site for each customer. Generating a sum tag value, and selecting sites belonging to an upper predetermined rank for each customer based on the sum tag value. The method of claim 9,
And the step (iv) multiplies corresponding tag values by different predetermined weights according to the log information tag for each site, and adds them. The method of claim 8,
The similarity value calculation between sites in the process (ii) is

Is calculated through, where
a = (a ₁ , a ₂ ,…, a _m ): tag values of a site in the site tag table
b = (b ₁ , b ₂ ,…, b _m ): Tag values of other sites in the site tag table
m is the total number of customers with site purchases. The method of claim 11,
The process (v) is based on the information of the predetermined rank higher sites selected through the process (iv) for each site and the sites having a similarity or more than a predetermined value obtained in the process (ii), Using the number of similar sites (Voting value) and the similarity value with the similar sites for each site of the selected top rank sites, a recommendation value (Score) is calculated, and the recommendation value And a predetermined number of sites are presented in the large order as the customer log customized recommendation site. The method of claim 12,
The (v) process may include a first site that is any one of the higher rank sites selected through the (iv) process, and one or more second sites that are similar to the first site. Based on the similarities, the total number of the second sites is calculated, a maximum value of the similarity values between the first site and the second sites is calculated, and then, for each of the second sites. And calculating a recommendation value for the customer. The method of claim 13,
The recommended value is the following formula

Where cosine is the largest of similarity values with related sites of a particular site, voting_weight is a weighting constant, and voting is a second, similar to the first site, wherein the similarity value is greater than zero. The number of sites, max (voting, each customer) is a server, characterized in that the maximum of each customer's voting value.

Images