KR20150066644A - Personal Action-Based Interest and Preference Analysis Method and System - Google Patents

Abstract

The present invention relates to a system for analyzing personal behavior-based preference and, more specifically, to a personal behavior-based objects of interest and preferring fields analysis system, which connects to a website of an operator with wired/wireless network with a personal terminal and understands objects of interest and preferring fields based on progressed transaction data, includes a personal behavior-based objects of interest and preferring fields analysis computer (server) which collects the transaction data in real time and displays personal characteristics related to products, the objects of interest, and the preferring fields from the highest ranking to the lowest ranking according to access frequencies per product, correlations related to purchases, and the latest representing the latest inquired product. In the correlation related to the purchase, the more certain connection location of the website of the operator associated with the purchases, the heavier weighting applied.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a personal action-based interest and preference analysis system,

The present invention relates to a system for analyzing a subject of interest and a preferred field (or area) of an individual based on personal behavior, and more particularly, to a system for analyzing a service (system) Based on the log data, the service utilization (system) utilization of the individual is determined, and the service usage (system) of the individual is observed in real time and cumulatively in terms of frequency, freshness, and relevance (or relationship) , Personal information such as individual's first name, gender, age, educational background, salary, etc. can be utilized at all, and the individual's interests and preferences can be analyzed regardless of service (system) To a system for analyzing personal behavior based interests and preferences.

Due to the development of wired / wireless communication networks and the development of personal terminals, many products such as books, home appliances, household goods, and other foods are purchased online rather than offline.

From the viewpoint of the operator or the seller who operates the site for such purchase activity, there is a great need to grasp the customer's preference.

On the other hand, there are many online purchases that consumers can overcome the limitations of time, and they are surfing the web simply by searching for specific items before purchasing.

Therefore, it is a reality that sellers should grasp, analyze and cope with consumers' consumption propensity or purchase propensity. In fact, most of these processes are sending advertisements for repurchase to customers who purchased specific goods by email or mail, In some cases, they sent e-mails or mail to those who would receive them. As a result, most advertisers are randomly indiscriminately advertising to an unidentified consumer in the case of an advertising company that needs to make promotions for sellers or sales.

When consumers look at the steps to purchase, it is common to go through the stages of interest to purchase.

Until now, however, consumers did not pay much attention to what they were interested in or what they were doing when they accessed certain websites.

Meanwhile, due to the development of computer-related information and communication technology, discussions about big data (BIG DATA) have been actively conducted in the related industries, but the discussion about this is insufficient.

On the other hand, as a smartphone becomes popular, a program such as facebook, a so-called social network program for communicating and exchanging messages with others on a smart phone, There is also a program that shows data (such as data associated with people on your friends list). However, this is the level of reasoning based on input data, simply showing the favorite material of people in the same group. Collaborative filtering is a sort of collapsing filtering that shows that the accuracy of associations between the data is low and that a friend likes what I like and that my friend clicks like (like in English) Although it may be useful for first-time visitors, if you are on my buddy list, you can show me what you like, but it's hard to say that it's meaningful information.

[Related Bibliographic Information]

1. System and method for providing product recommendation service (Patent Application No.: 10-2008-0049457)

2. Purchasing Pattern-Based Shopping System and Method (Patent Application No. 10-2011-0005275)

An object of the present invention is to provide a system for analyzing a subject of interest and a preference field of an individual based on personal behavior, and more particularly to a system for analyzing a service (system) use log data (System) usage history of a corresponding individual on the basis of the frequency, the latestness, and the relevance (or relationship) of the individual, (Personal information) such as name, gender, age, educational background, salary and so on, and analyzing the individual's interests and preferences regardless of service (system) And to provide a personal behavior based interest and preference analysis system.

According to a preferred embodiment of the present invention to achieve the above object, a personal behavior based interest and preference analysis system comprises:

1. A personal interest based interest and preference analysis system for accessing an operator web site via a wired / wireless communication network through one or more personal terminals and determining preference based on the transaction data processed,

The transaction data is collected in real time, and the frequency of accesses for each session is stored and calculated. The relationship with the personal terminal data accessed on the corresponding operator web site is displayed as the first display data based on the calculated frequency, 1 display data is divided into specific sessions and aggregated in real time to display the relationship with the customer's data as second display data, and the second display data is aggregated by the specific session on a day-by-day or month-by-month basis, And a personal behavior based interest and preference analysis computer (server) that divides the relationship into third display data and displays it.

In addition, the first display data is data for displaying a specific customer, a specific product, and a specific product access frequency of a specific customer. In the first display data, at least one specific product corresponding to a specific customer is displayed, The data indicating the specific product connection frequency of the specific customer according to the number of the specific products.

The second display data is data indicating a specific customer, an item of interest (target or an interested item) and a score, and the score of the second display data is a score calculated by applying frequency, relevance, and freshness. (Frequency) calculates the number of inquiries of the same commodity. Relativity is calculated by applying a higher weight to the purchase, assuming that product interest is different for each route and action. And the total score is calculated by calculating the number of times that the interest of the commodity is higher and the latest inquiry item is selected in the case of the same score.

In the second display data, at least one number of interested objects corresponding to one specific customer is displayed, and the score of the second display data calculated for each of the interested objects is displayed. The total score of the number of interested objects The high customer is displayed first.

The third display data is characterized by displaying a specific customer, a preferred product category, and a score. The score of the third display data is a score calculated by applying frequency (number), freshness, and relevance, and frequency Frequency refers to the number of items of similar interest for one month, assuming that the preference of the items of interest is higher. Relativity is the highest ranking item, And the recency is calculated by summing up the scores calculated by calculating the preference of the latest interested items in the case of the same score, assuming that the preferences of the recent interested subjects are higher .

In addition, the score of the third display data represents the customer's preference category, the customer's preference category is one or more, and the score of the third display data displayed in the preference category of the customer is higher And displays the image.

Preferably, the personal behavior based interest and preferred area analysis computer (server) communicates with the personal terminal through a wired / wireless communication network or directly through a communication module in a personal terminal,

Data collection means for collecting service (system) utilization log data,

A user session determination means for determining start and end of a session for dividing a process of collecting usage log data by one session of a user,

FRR model applying means for analyzing the interest applying the FRR model based on Frequency, Purchase Relativity, and Recency to the usage log data collected by the user's session,

Based on the data obtained by the FRR model applying means for the analysis of interest, the interest of interest based on the individual behavior (including not only the product but also the object related to the user's use behavior) is extracted A display means for displaying interest from the highest ranking to the lowest ranking,

 FRR model application method for analysis of preference field applying FRR model based on frequency, purchase, relativity, and recency to the interest ranking and score data collected by each user session,

The data obtained by the FRR model applying means for analyzing the preference field may be used as a preference field based on the personal behavior (the attribute information held by the object information related to the user's use behavior is not limited to the category or color of the product, And a preference field display means for displaying the preferences for preferences of the preference degree from the highest priority order to the lowest priority order.

According to another aspect of the present invention, there is provided a system and method for analyzing a user's interest and preference based on personal behavior based on input usage log data, the system comprising: And displays the relationship with the personal data connected to the program on the basis of the calculated frequency as the first display data. The first display data is classified by the specific session and is aggregated in real time to display the second display And a personal behavior based interest and preference analysis program for displaying the second display data displayed in the form of data, aggregating the displayed second display data for each specific session on a day-by-month or month-by-month basis, A personal behavior based interest and preference analysis system is provided.

According to another aspect of the present invention, there is provided a method for analyzing a user's interest and preference field based on personal behavior using a personal behavior based interest object and preference field analysis system, the method comprising: collecting and displaying usage log data of a user A second step of determining the end of the session of the user, a third step of storing basic information of the user based on the usage log data, A fourth step of generating and displaying a list of interest through the information calculated by adding up the frequency and the number of frequencies, a fifth step of categorizing and storing the list of interest into sessions, , And the information calculated by adding the frequency is displayed together with the attribute information, This field indicates the preferred target the individual behaviors comprising the step 5, based interested displaying the list and the favorite field analysis method is provided.

According to another aspect of the present invention, there is provided a personal behavior-based interest object which is used in a specific terminal and is used to identify a target of interest and a favorite field based on usage log data generated when a service (program) A terminal having a preference field analysis program, the terminal collects the transaction data in real time, stores and calculates the frequency of accesses for each session, and based on the calculated frequency, The second display data is aggregated on a per-session basis or on a monthly basis to display the second display data on a per- Based on the personal behavior that displays the relationship between the first and third display data A terminal capable of analyzing a personal behavior based interest object and a preference field is provided.

As described above, according to the personal behavior based interest and preference analysis system according to the present invention, it is possible to provide a system and method for analyzing a user's interest based on a service (system) ) The utilization history is observed in real time and cumulatively from the viewpoint of frequency (number), relevance and freshness, and the interested object and the preferred area of the individual are analyzed and utilized for sales, marketing, and services.

In particular, since the present invention does not utilize personal information such as name, gender, age, salary, and address of an individual at all, it has high personal information security, (System) of the service (system) of the individual to be analyzed, and not the statistical analysis method that utilizes the data of the customer of the whole service (system) The amount of data to be analyzed is small because only the usage log data, which is data generated when the service program is used in the corresponding terminal, operates only in the terminal.

1 is a schematic block diagram of a personal behavior based interest and preference analysis system according to the present invention.
2 is a schematic diagram of a terminal having a built-in personal behavior based interest and preference analysis program according to the present invention.
FIG. 3 is a block diagram showing a concrete configuration of a personal behavior based interest object and preference field analysis computer (server) of FIG. 1; FIG.
FIGS. 4 to 6 are diagrams schematically illustrating a basic information processing process of a personal behavior based interest object and preference field analysis system.
FIGS. 7 and 8 are explanatory views schematically illustrating an example of a personal information processing process of the personal behavior based preference analyzing system according to the present invention.
9 is a schematic flowchart of a personal behavior based preference analysis system according to the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a personal behavior based interest and preference analysis system according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

Referring to FIG. 1, the Personal Behavior Based Interest and Favorite Sector Analysis System according to the present invention can be used for analyzing a significant interest and preference field quickly in a mass data processing environment from a multi-source system, The system is designed to provide the user with a preference that appears in the user's personal behavior, the interest in the user's personal use behavior, and the repetitive interest in the user's personal interest. It is a CIP analytical model that analyzes the time series.

In the present invention, analysis of C, I, P, i.e., characteristics, interests, and preferences is performed logically based on data. In other words, FRR model is used when extracting interest from information according to the characteristics of user's personal use behavior. The FRR model will be described later.

On the other hand, the service (system) utilization log data required to generate the characteristic data includes 1) personal identification information, 2) time information on which the individual has used the activity, 3) object information related to the usage activity, 4) And includes basic information necessary for obtaining history information as characters and symbols. In the present invention, by utilizing such basic information, the use behavior details of the user are obtained and the usage behavior characteristic data of the user is generated.

1) In the case of personal identification information, when the service (system) is being used in the state of user identification (user login), the user identification code value defined by the service (system) is used as identification information of the individual, (Non-log-in) state, an identification code value that can distinguish the terminal (terminal device of personal PC, smart phone, tablet PC, etc.) permanently using the service (system) or the service (system) is used.

4) In the case of detailed usage activity information, it refers to concrete contents using the service (system), and it is a concrete example such as "turn on VOD video", "turn on music", "place product in shopping cart".

3) In the case of the target information related to the use activity, in the specific example of the detailed use activity history information "Turn on the VOD image", the specific VOD video material that the user has turned on such as "Infinite Challenge 25 times" The VOD video identification code value defined by the service (system) is used to distinguish the video object.

Also, in the 'target of interest' and 'preference field' of the individual analyzed by the present system, the 'target of interest' is represented by the target information related to the use of 3) If the system is applied to the VOD video service (system) , The subject of interest analyzed is represented by each VOD video object such as 'Infinite Challenge' or 'Year of the Sun'. In this case, the 'preference field' is expressed as a 'preference field' in terms of attributes possessed by the VOD video, ie, genre, actor, and release date. For example, the genre 'preferred field' is represented by 'drama', actor 'preferred field' by Kang Ho Dong, and the release date 'preferred field' is expressed by 'released within the last 3 months'.

On the other hand, in the present invention, attention data (hereinafter referred to as second display data) is extracted from data extracted from characteristic data (hereinafter referred to as first display data), and preference data (hereinafter referred to as third display data And a ranking list is displayed. The description thereof will be continued below.

The personal behavior based interest and preference analysis system according to the present invention collects the personal behavior based interest and preference domain through a wired / wireless communication network through one or more personal terminals P1 to Pn (Or a program) collected directly through a personal behavior based interest and preference analysis program embedded in a personal terminal within the corresponding personal terminal, Preference field). At this time, the data collected by the user is independently processed and processed as data on the individual's interests and preferences of the individual based on the behavior of the individual, wherein the data of the different users do not affect each other but are displayed on the list .

In the personal behavior based interest object and preference region analysis system according to the present invention,

In an example applied to the merchandise sales service, the collected usage log data is collected in real time, and the frequency (frequency), (purchase) relativity, and the recency indicating the latest inquiry item And a personal computer (server) 100 for analyzing a favorite subject and a favorite region based on personal behavior, which displays the characteristics of the personal use activity, the interested goods, and the preferred product category information in descending order.

2, the analysis of the object of interest and the preferred region may be performed in the individual terminal. In this case, the personal behavior-based interest and preference analysis program 100-1 may be used for a specific terminal, for example, Tabs, notebooks, and the like. In other words, a personal behavior based interest and preference analysis program 100-1, which is embedded in the specific terminal and is used to identify interest objects and preference fields based on usage log data generated when the service program is used, ) Stores transaction data on a terminal in real time, stores and calculates the frequency of accesses on a session basis, and calculates a relationship between personal data that is connected to the program based on the calculated frequency, The second display data is aggregated on a per-session basis or on a monthly basis to display the second display data on a per- Based on the personal behavior-based interest and preference analysis program It characterized in that it comprises a. In this case, the function may be performed in the same manner as performed through the personal behavior based interest and preference analysis computer server. These functions are the same as those described in Figs. 1 to 6, and therefore, the following description will be referred to for the detailed description.

On the other hand, the above-mentioned (purchase) relativity means that the higher the relevance of the detailed use activity history is with the target of the service, for example, in the case of a shopping mall, And the higher the weight is applied, the more relevant it is.

The personal behavior based interest and preference domain analysis computer (server) 100

Communication is performed with the personal terminals P1 to Pn through the wired / wireless communication network or directly through the communication module 110 and the control unit 120 in the private terminal

Data collection means for collecting service (system) utilization log data,

A user session determination means for determining start and end of a session for dividing a process of collecting usage log data by one session of a user,

FRR model applying means for analyzing the interest applying the FRR model based on Frequency, Purchase Relativity, and Recency to the usage log data collected by the user's session,

Based on the data obtained by the FRR model applying means for the analysis of interest, the interest of interest based on the individual behavior (including not only the product but also the object related to the user's use behavior) is extracted A display means for displaying interest from the highest ranking to the lowest ranking,

 FRR model application means for analysis of preference fields applying FRR model based on Frequency, Purchase Relativity, and Recency to the interest ranking and score data collected by user's session,

The data obtained by the FRR model applying means for analyzing the preference field may be used as a preference field based on the personal behavior (the attribute information held by the object information related to the user's use behavior is not limited to the category or color of the product, And a preference field display means for displaying a preference display order from the preference order to the lowest preference order in order of preference.

The control unit 120 controls the entire apparatus, and manages and stores login data, temporary storage of data, connection with other web servers, temporary data storage, etc., for personal behavior based interest and preference analysis of the present invention Processes the data of the entire process, and outputs a function control signal corresponding to a specific input signal. The general functions of the controller 120 are similar to those of a conventional computer control unit (a CPU including a storage memory of a kind), and thus a detailed description thereof will be omitted, and the contents of the present invention will be mainly described.

The FRR model applying means may be internally divided into a frequency (frequency) model applying means, a R (relativity) model applying means, and an R (recency) model applying means. For example, F model applying means may be initially applied, and then RR model applying means may be applied simultaneously or sequentially.

In the present invention, ranking and score information are generated for each of I and P representing a target object and a favorite area.

The most important factor in the present invention is the frequency.

In addition, additional recency is important. That is, the higher the currentness, the higher the weight. The latest detected commodity code has a higher score.

In addition, relativity is important. When applied to shopping services, relevance is purchase relevance. If the location where the individual is currently browsing through the personal terminal (viewing is one of the concrete usage activities) is closer to the purchasing stage in the shopping activity progress stage, the weight becomes higher. That is, since the personal data containing the goods in the shopping basket among the goods information in the main screen and the goods stored in the shopping basket are closer to the purchasing process, they are processed (or calculated) with a high weight.

The frequency, relevance, and freshness are referred to as the FRR model by citing the first letter of each alphabet, and the frequency is applied first, and then the relevance and the renewal can be sequentially applied. However, the present invention is not limited to this, and it is possible to apply the renewal after applying the frequency, and thereafter, the relevance can be grasped and applied.

As in the case of daily life, if interest continues on a particular object, it can be considered to be liked. (Eg, goods, contents, etc.) data based on an individual's behavior, that is, related to an individual's current use behavior.

For example, if a commodity code for a commodity that the individual clicks on and connects to a specific site is continuously detected, data on the commodity is continuously processed, and ranking and scoring information are generated. . This list of interest is cumulatively managed. (The target list of interest for each individual session is stored and managed as rank and score information).

Therefore, in the present invention, it is possible to process big data in real time without any technical help or method for processing big data.

In the present invention, when a person recognizes which object shows interest, it can be grasped through interest ranking information indicating that the interest ranking score is high. As a matter of course, it is a preference stage that is more severe than the stage of interest in everyday life. Therefore, the preference ranking and the score information are summarized as a factor of preference.

Preferences If the subject of the survey is any service (system) field, such as a sound source (or music), a commodity, a public opinion survey, etc., the personal use behavior characteristic, interest ranking and score, Rank and score information.

In the past, the consumer's preference for the product was difficult to apply to other objects because the object of the survey was a product, or the operator made a predetermined inquiry or a predetermined method. However, in the present invention, Whether the service is a system (service), it can be easily grasped step by step based on frequency, relevance, and up-to-date.

That is, the analysis of the object of interest and the preferred area in the present invention is different from any existing method in this method. In the present invention, the preferred region is identified step by step in real time based on an individual's behavior. First, we identify the characteristics of the use behavior, second, identify the target of interest, and finally, identify the preferred area. At this time, based on individual service (system) utilization log data, it is judged and calculated, and rank and score of each step are displayed.

In the present invention, the difference between grasping the object of interest and the preferred region will be described. For example, if an individual is interested in a Samsung air conditioner model 1 with a price of one million won and a Samsung air conditioner model C with a price of 1,110,000 won, the preferred category for the product category is air conditioner or seasonal household appliances, As the manufacturer's preference area is determined by Samsung Electronics at 1.1million Won, there is a difference between the subject of interest and the preference area in this respect. For example, if the song A, the song B, and the song C of the music service are interested in the top rank, the preference object means the product level, i.e., the target information level related to the user's use behavior. It is possible to know the presence of Cho Yong Pil in the top ranking of the singer preference area of the person by the preference area grasp.

For reference, if you look at the ranking information in the ranking and score of each step, the latest one is ranked higher.

INDUSTRIAL APPLICABILITY The present invention can be applied to various service (system) fields such as music service, commodity sales service, and image providing service. The C-I-P and the frequency, the recency, and the relativity determined according to the present invention are capable of processing real-time input data, and are highly accurate in recognizing an object of interest and a preferred area of an individual.

In the present invention, preference can be extracted by using personal information. However, in the present invention, personal information such as age, sex, race, name, education, property, address, Interests and preferences can be known.

FIGS. 4 to 6 are diagrams schematically illustrating a basic information processing process of a personal behavior based interest and preference domain analysis system according to a preferred embodiment of the present invention.

Referring to FIGS. 4 to 6, the personal behavior based interest and preference analysis system may be configured to analyze the interest and preference domain based on the service (system) utilization log data generated in the process of using a specific service (system) Identify the service (system) usage history and identify the subject of interest and preference based on the individual's service (system) usage history.

In the personal behavior based interest and preference analysis system according to the present invention, first, the service utilization log data of the individual is collected and stored in a session (a series of action units from the start to the end of service use) And stores the collected data in a time sequence. The collected data is stored in a personal database, which is stored in a personal database The specific use behavior history information and the target information related to the use behavior are listed and stored together in the order of the time information in which the individual used the action. The basic data related to the service use is displayed as the first display data for each session of the service (system) user thus sorted, the first display data is divided for each session, and the target information related to the use behavior is listed as the identification information of the individual The score is calculated by considering frequency (frequency) and (purchasing) relativity, that is, specific usage activity information, for each target information related to each usage activity. If the score is the same in the target of interest, the target information related to the recency, that is, the recent usage, is counted as a higher priority.

Items of basic information shown in Fig. 4 include individual identification information, time information on an individual's use behavior, object information related to a use behavior, and specific usage behavior history information.

Examples of specific usage activity history information include 'start service use', 'view product description', 'view product purchase history', 'put product into shopping cart', 'terminate service use' .

The contents of the first display data classified by the session include items such as 'time information on personal use', 'time information on personal use', and 'time information on personal use' in the session of user U1 from 13:24 to 13:30 on November 1, Subject information related to the use behavior, " and " specific usage behavior history information " may be displayed.

The target information list and the scoring data related to the ranking order use operation in descending order are displayed as the second display data for each session of the service (system) users thus sorted, and the second display data is used as the identification information of the individual based on the displayed second display data (Frequency) and (purchasing) relativity score for each attribute information, that is, ranking and score information in the interested list are considered, It compiles rankings and scores by field list. If the scores in the list of preference fields are the same, the attribute information held by the object information related to the latest usage (ie, recent usage) is counted as a higher priority.

5, in the information on the session from 13:24 to 13:30 on November 1, 2013 of the user U1, the relevance score (example) for the specific use activity history information displayed on the lower left side, , Which includes specific usage behavior details and (purchasing) relevance score items, which have different score items. For example, you could differentiate scores by item to give (purchase) relevance scores of 0.2, 0.4, 0.6, 0.8, 1.0, and 1.2.

In addition, the contents displayed on the upper left side of FIG. 5 show the target information related to the use behavior, the frequency, the relevance score in consideration of the detailed use behavior details, and the subtotal (example) items. The score of the relevance score and the frequency score considered is displayed. 5, the second display data as the second display data includes the interest ranking, the target information related to the usage behavior, and the sum (example) item. The second display data for each session of the user U1 is generated.

The ranked list of preferred fields and the score data for each attribution information held by the target information related to the use behavior of each service (system) user classified by the third display data are displayed separately.

Referring to FIG. 6, a process of generating third display data for each attribute information held by the target information related to the use behavior of the user U1 is shown. That is, a list and a score of interest ranking of the user U1 per session (i.e., the second display data in FIG. 4) are displayed from the past session to the latest session. Here, the attribute information possessed by the target information related to the use action is displayed. The attribute information displays a category, a color, and the like as target information and attribute information (for example) related to the use behavior. In addition, (Example) item, which includes a ranking item of interest and a (purchasing) relevance score item.

The data displayed in the upper right part of FIG. 6 includes items of interest information, frequency, interest ranking, subtotal (example), attribute, category, attribute, color, . The third display data item is displayed in the lower right portion of FIG. 6, and the third display data item includes an attribute; category preference field ranking; attribute; category preference field; and sum (example) Color preference domain ranking, attributes, color preference domain, and sum (example) items. If the total score is the same in the preference field ranking, the attribute information held by the object information related to the recent use behavior is counted as a higher priority. In addition, the third display data may be expanded to a variety of attribute-based preference categories held in the target information related to the use behavior in addition to categories and colors.

Figures 7 and 8 are illustrative illustrations of a personal behavior based interest and preference domain analysis system in accordance with the present invention.

Referring to FIGS. 7 and 8, it is possible to quickly grasp characteristics of significant usage behavior in a large-capacity data processing environment of a multi-source system, which is an important matter for a personalized marketing solution, A CIP analysis model has been provided for analyzing the characteristics, interests and preferences of the users.

The customer A displayed on the left side of FIG. 7 accesses a specific web site and performs a variety of service utilization activities. The usage log data generated in the usage process passes through a wired / wireless communication network or internally through a communication module, And processed through an object and preference region analysis computer (server) 100.

When a user makes use of a service (system), the characteristic of the use behavior is analyzed and displayed in a characteristic item. In the second stage, the information of interest is identified by ranking and score. In the third step, it is judged that the repeated interest is a preference, and the preference region information is ranked and indicated by the score. Frequency, Relativity, and Recency data are used in each step.

The frequency, frequency, and frequency of use of the service (system information) related to the corresponding product (inquiry of product information, product (Ie, behavior included in shopping activities such as shopping, pouring, putting in a shopping cart), and the recency refers to the number of items of interest related to recently used usage high. Also, relativity is higher than the interest index of the use behavior that is highly related to purchasing. In the end, interest ranking and score information are displayed from the highest rank to the lowest rank based on these three variables.

In addition, as to frequency, freshness, and relevance used to indicate the preference region according to the product category attributes as ranking and score information, frequency is the number of times the corresponding commodity is displayed in the interest ranking, Recently, the preference index of the product category of the recently generated interest goods is higher than the preference index of the commodity category of the same interest goods that occurred in the past. Relativity is also higher than the product category preference index of the top-ranked interested products that are highly related to purchasing. Eventually, the ranking and score information of the preferred areas of each product category attribute are displayed from the highest rank to the lowest rank based on these three variables.

Referring to FIG. 8, a process of inferring a target interest category and a preference category of a shopping mall customer based on the CIP analysis model is exemplarily illustrated. The FRR (frequency / relevance / renewal) Interest lists and customer preference categories are analyzed and updated.

As shown, the initial time series transaction can be divided into a hit (log), a purging log, and a purchase log. However, the present invention is not limited thereto.

8, the service (system) utilization data is aggregated in real time. Thus, it becomes possible to compile data that can identify the characteristic (C) of the customer's use behavior. The customer (denoted as V), the commodity (denoted as P), C denotes the characteristic and CC denotes the product category.

That is, in the present invention, it is performed in the order of the service (system) use data-> C: the characteristics of the customer's use behavior-> I: the list of items of interest to customers-> P: Data is aggregated in real time in steps C and I, and when I performs P in I, a period is determined and periodically aggregated to display, for example, one month of data (including, but not limited to, daily data may be displayed ).

In FIG. 8, the CIP model is described. The characteristic C (Character) represents a statistical aggregation of characteristics of a customer's use behavior. Also, interest I (Interest) is an object that continues to appear in the ranking result. Also, the preference P (Preference) is an object that is repeatedly observed among the objects of interest.

Frequency, number of frequencies, relevance and frequency used to express the interest in the FRR model for the analysis of interest will be described in detail. Frequency is the frequency of frequently viewed items The higher the standard, the greater the number of inquiries. Also, relativity is based on the fact that product interest is different according to use behavior (for example, steam / inquiry). For example, the weight of the product is higher than that of the simple product information. In addition, the recency is based on the fact that the inquiry item is more interested in the recent inquiry item, and the calculation item (calculation) is the item with the same item.

In addition, in the FRR model for preference domain analysis, frequency (number), relevance, and frequency used to indicate the preference of the product category are expressed in frequency. The criterion is applied that the preference of the product category is higher, and the calculation scores the number of the cumulative period of the same interest category. Relativity is calculated by applying a higher preference to the category of interest in the top ranking and applying a higher weight to the top ranking category of interest. Also, the recency applies the criterion that the preference of the interested product is higher recently, and the calculation (calculation) selects the latest interested product in the case of the same score.

In FIG. 8, V1 and V2 represent different customers (or users), and the C data of V1, the I data and the P data are automatically sorted from right to left in the order of scores. In the case of the C data, which is displayed horizontally long for the V1 customer, it is shown from the left to the right in the order of the number of times. In the case of the I data, the score for the interested object is high. Are shown from the point of high score in the preferred field. Data of customers V1 and V2 are displayed separately from each other and displayed only in the administrator mode, and are displayed in the same manner with respect to customer V2. C data, I data, and P data Are automatically counted and displayed in order.

Figure 9 is a schematic flow diagram of a personal behavior based interest and preference analysis system in accordance with the present invention.

Referring to FIG. 9, in the personal behavior based interest and preference analysis system according to the present invention, when an individual is connected to a specific site through a terminal (including a smart phone) connected to the Internet via wired or wireless, The behavior based interest and preference domain analysis computer (server) 100 is collected (ST-10). At this time, the data collection is processed by the data collection means. Then, the user's session determination means determines the end of the session (ST-12). On the other hand, the data collected at the step ST-10 is feature data and the upper ranking is displayed based on the frequency of the interested object of the customer. When there are a plurality of items of interest of the customer, . In the case of multiple products, the products with the highest frequency are displayed in order from left to right.

The FRR model is then applied through the FRR model applying means for the analysis of interest (ST-14). The FRR model applying means for the analysis of interest can process the collected data (transaction data) by first judging only the frequency. It can also process collected data (transaction data) applying both frequency, relevance, and freshness. The data collected in the first session is the most important factor, and the relevance can be zero if it is numerically calculated (if there is no purchase relevance, the value itself is zero and the weight is zero) They are all the same numerical values, or 0, since they are all the latest data at the beginning). Therefore, the first session data is the most important data and the data of the subsequent session is the most important data. Therefore, the FRR model applying means for analyzing the interested object first determines the first session data. In subsequent sessions, the relevance and freshness data will act as a kind of data correction.

The data processed by the FRR model application means for the analysis of interest is generated and displayed by the interested display means in order from the highest ranking order to the lowest ranking order (ST-16). There is no limitation on the kind of data displayed by such a target display means. That is, the top list displayed by the interested target display means indicates the degree of interest of a particular movie, for example, a melody, an action, a drama, etc., when the connected person accesses the movie introduction site. The list of items or objects of interest may be more than one as an example of interest displayed to the customer.

On the other hand, the top list displayed by the interested display means (that is, the interest list of the customer, the interest item of the customer in the case of the commodity, or specifically the interested commodity list) ) The FRR model application method for analyzing the preference domain is applied again (ST-20). In this way, the user can display his / her favorite area beyond the target of interest (ST-22). At this time, the preference area includes the preferred product category for each customer, and the score of each category (the score calculated by applying the FRR model for analyzing the preference area) is displayed together with the ranking. There may be more than one customer preference category. In addition, the items with high scores of interest for a specific customer are displayed in order from left to right.

That is, FRR model application means for analyzing the preference region (or preference field) is applied to the interest top list, the number of times of detection is calculated in the top list by applying the frequency F first,

We then apply relevance R to apply weights, calculate relevance,

Then, apply the latest R, but if the scores are the same, the more recent ones are weighted.

By applying the FRR model applying means for analyzing the preference region to the I list which is the list of interest, the P list which is a list of preferred regions is displayed by the preference region displaying means. Here, a plurality of P lists may exist. Also, if there are multiple customer preference categories, the highest preference category with the highest score is displayed first from left to right.

As described above, according to the personal behavior based interest and preference region analyzing system according to the present invention, it is possible to provide a personal and personal preference analyzing system for analyzing an individual who is transmitted through a specific personal terminal or personal PC or tablet terminal on a specific web site or a wired / (Transmission) based on the behavior of an individual using a characteristic service (program) on a screen (information) screen is divided into characteristics (characteristic), interest (degree), and preference Based on the data obtained by the frequency of the first customer's specific interest, the latest and the relevance (or relationship) are weighted and scored based on the frequency, and the top list is calculated based on the score Based on the behavior of the individual, regardless of the field (applicable to art, music, commodities and all other fields) Phosphorus can analyze the preferred areas of interest beyond the interest and the effect that can be utilized such as purchasing, marketing through it.

In the embodiment of the present invention, the first characteristic data is obtained for the customer based on the frequency, and the top list is displayed. However, the present invention is not limited to this, and the top list may be displayed by generating data based on the frequency and the relevance. The location on the web site is different for each purchased product and the code (for example, product code) is different, so that the data can be collected and used by the operator computer (server).

110: communication module
120:
130: Analyzing program of interest and preference area

Claims (32)

A personal behavior based interest and preference analysis system for accessing an operator web site via a wired / wireless communication network through one or more personal terminals and determining preference based on the transaction data processed,
The transaction data is collected in real time, and the frequency of accesses for each session is stored and calculated. Then, based on the calculated frequency, the relationship with the personal terminal customer data accessed on the corresponding operator web site is displayed as the first display data, The first display data is classified by a specific session and aggregated in real time to display the relationship with the customer data as the second display data and the second display data is aggregated by the specific session on a daily or monthly basis, A personal behavior based interest and preference analysis computer (server) for displaying the personalized behavior based on the first display data and the third displayed data. The method according to claim 1,
Wherein the first display data is data representing a specific customer, a specific product, and a specific product access frequency of a specific customer. 3. The method of claim 2,
Wherein the first display data is data for displaying one or more specific products corresponding to one specific customer and displaying the specific product access frequency of the specific customer according to the number of the characteristic products from a high rank. Behavior based interest and preference analysis system. The method according to claim 1,
And the second display data is indicative of a particular customer, item of interest (object or item of interest), score, and a personal behavior based interest and preference analysis system. 5. The method of claim 4,
The score of the second display data is the score calculated by applying the frequency, relevance and freshness. Frequency is the frequency of inquiry of the same product. Relativity is calculated by multiplying the product interest by path and action The higher the recency is, the higher the interest rate of recently inquired products. In the case of the same score, the most recent recycled items are selected first, Wherein the personalized behavior based interest and preference analysis system comprises: 5. The method of claim 4,
The second display data may include at least one item of interest goods corresponding to a specific customer, and may display the score of the second display data calculated for each item of interest, Wherein the first and second preferences are displayed first. The method according to claim 1,
And the third display data displays a specific customer, a preferred product category, and a score. 8. The method of claim 7,
The score of the third display data is calculated by applying the frequency (number), the freshness, and the relevance, and the frequency (frequency) is the number of items of interest And the relativity is calculated by applying a higher weight to the top-ranked items of interest, assuming that the preference of the items of interest is higher in the top rankings, and the recency is calculated by adding more recent items of interest And the total score is calculated by calculating the number of times that the highest interest item is selected first in the case of the same score. 9. The method of claim 8,
The score of the third display data indicates the customer's preferred product category, the customer's preferred product category is one or more, and the score of the third display data displayed in the preferred product category of the customer is displayed from the rank with the highest total score Based personal behavior based interest and preference analysis system. [2] The method of claim 1, wherein the personal behavior based interest and preference region analyzing computer (server) communicates with the personal terminal through a wired / wireless communication network or directly through a communication module in a personal terminal,
Data collection means for collecting service (system) utilization log data,
A user-by-user session determination means for determining start and end of a session for dividing a process of collecting usage log data into sessions,
FRR model applying means for analyzing the interest applying the FRR model based on Frequency, Purchase Relativity, and Recency to the log data collected for each session,
Based on the data obtained by the FRR model applying means for the analysis of interest, the interest of interest based on the individual behavior (including not only the product but also the object related to the user's use behavior) is extracted A display means for displaying interest from the highest ranking to the lowest ranking,
FRR model application method for analysis of preference field applying FRR model based on Frequency, (Purchase) Relativity, and Recency to interest ranking and score data collected by user session,
The data obtained by the FRR model applying means for analyzing the preference field may be used as a preference field based on the personal behavior (the attribute information held by the object information related to the user's use behavior is not limited to the category or color of the product, And a preference field display means for displaying a preference display means for displaying a preference for a preference degree from a preference order to a preference order display order, And preference analysis system. The service system use log data generated during the process of the user using the operation service system of the specific operator through the terminal is collected. The collected usage log data of the collected user is classified and analyzed according to the session, And a preference field based on the user's continued use of the service (system). 12. The method according to claim 11, further comprising the step of collecting the personal service system use log data and analyzing the personal service system use log data according to the frequency of the subject information related to the use behavior generated in the session in the session And the specific use activity history information are collected and stored, and a series of use activities of the individual is organized in time order for each session, and the organized data is classified into specific usage activity history information and corresponding usage activity information The basic data related to the service use is displayed as the first display data for each session of the sorted service system, and the first display data is divided for each session, (Frequency) and (purchase) information for each target information related to each use activity ) Relevance information is taken into account and the scores are aggregated and displayed. 12. The method according to claim 11, wherein if the scores in the order of interest are the same, the recency, i.e., the object information related to the recent use behavior is aggregated as a higher priority, system. 12. The information processing apparatus according to claim 11, wherein the utilization log data includes identification information of an individual, time information on an individual's use behavior, object information related to a use behavior, and basic information including specific usage history information Personal behavior based interest and preference analysis system. 12. The method of claim 11,
The contents of the first display data classified by session include items that include time information on an individual's use behavior, object information related to a use behavior, and specific use behavior history information in a session for a specific period of the user U1 as an item A personal behavior based interest and preference analysis system characterized by. 12. The method according to claim 11, wherein the object of interest is second display data in which data calculated by adding a frequency (number) to a relevance score for each user's detailed usage activity history information sorted by session is sorted in descending order, And displaying the target information and the interest ranking and total score related to the personal behavior based on the personal behavior in a descending order. [17] The method of claim 16, wherein the preference field is determined based on the second display data by searching for the attribute information held by the object information related to the use activity as the individual identification information, Purchase) relativity scores of the first display data and the third display data displayed in descending order in a list of preferred fields. 18. The method of claim 17,
When the scores in the preference field list are the same, the attribute information held by the object information related to the recent use behavior is counted as a higher priority, based on personal behavior based on the individual behavior based on interest and preference analysis system. A personal behavior based interest and preference analysis method using a personal behavior based interest and preference analysis system,
A first step of collecting and displaying usage log data of a user connected to a specific site,
A second step of determining the end of the session of the user,
A third step of storing basic information of the user based on the session log data,
A fourth step of generating and displaying a list of interest through information obtained by adding scores to the stored data and summing the frequency (number)
A fifth step of categorizing and storing the list of interest,
A fifth step of displaying a list of preference fields representing the preference order by property and a preference field by property by displaying scores of the stored interest list and displaying the calculated information together with the attribute information together with the frequency, Wherein the personal behavior based interest and preference analysis method comprises: 20. The method of claim 19,
The usage log data in the first step includes latest information including the identification information of the user of the user, the time information of the user, the target information related to the usage behavior, and the detailed usage history information, Wherein the personal behavior based interest and preference analysis method comprises the steps of: 20. The method of claim 19,
Wherein the ranking of interest is displayed in a list of interest in the fourth step, the ranking being indicated by a sum of the frequency of use of the user and the (purchasing) relevance score, And preference analysis methods. 20. The method of claim 19,
In the fifth step, the preference list is displayed in descending order from the highest ranking to the highest ranking in the interest list by summing the interest ranking items and the relevance scores for each level of interest indicated by the (purchase) relevance score in the list of interest Wherein the attributes of the preference field are displayed together. A terminal that is embedded in a specific terminal and has a built-in personal behavior based interest and preference analysis program that grasps interest objects and preference fields based on usage log data generated when a service (program) As a result,
The transaction data is collected in real time, and the frequency of accesses is stored and calculated for each session. Based on the calculated frequency, the relationship with the personal data connected to the program is displayed as the first display data, The second display data is aggregated in a specific session by day or month, and the relationship with the data per customer is divided into third display data and displayed Wherein the personal behavior based interest and preference analysis programs include personal behavior based interest and preference analysis programs. 24. The method of claim 23,
Wherein the first display data is data representing user data and a user's specific product connection frequency, wherein the user can perform an analysis of the user's interest and preference field based on personal behavior. 24. The method of claim 23,
Wherein the first display data is data for displaying at least one specific product corresponding to the user and displaying the specific product access frequency of the specific customer according to the number of the characteristic goods from a higher rank. And terminals capable of analyzing preferences. 24. The method of claim 23,
And the second display data displays a user, an item of interest (target or interested item), and a score. 24. The method of claim 23,
The score of the second display data is the score calculated by applying the frequency, relevance and freshness. Frequency is the frequency of inquiry of the same product. Relativity is calculated by multiplying the product interest by path and action The higher the score is, the higher the interest rate of the latest item is. In the case of the same item, the higher the score, the higher the score is calculated. Wherein the terminal is capable of analyzing a target of interest and a preference field based on personal behavior. 24. The method of claim 23,
Wherein the second display data is indicative of one or more of the number of items of interest corresponding to the user and the number of the second display data calculated for each of the items of interest is displayed. This possible terminal. 24. The method of claim 23,
And the third display data displays the user, the preferred product category, and the score. 24. The method of claim 23,
The score of the third display data is calculated by applying the frequency (number), the freshness, and the relevance, and the frequency (frequency) is the frequency of the same interest The number of items is calculated, and relativity is calculated by applying a higher weight to the top-ranked items of interest, assuming that the preferences of the items of interest are higher in the top rankings. And the total number of the calculated scores calculated by calculating the number of points to be used to select the latest interested items in the case of the same score. 24. The method of claim 23,
The score of the third display data indicates the user's preference product category, the user's preference product category is one or more, and the score of the third display data displayed in the user preference product category is displayed in the order from the highest ranking to the highest score A terminal capable of analyzing interests and preferences based on personal behavior. 24. The method of claim 23, wherein the personal behavior based interest and preference region analysis program comprises:
Data collection means for collecting service (system) utilization log data,
A user-by-user session determination means for determining start and end of a session for dividing a process of collecting usage log data into one session for each user,
FRR model applying means for analyzing the interest applying the FRR model based on frequency, purchase, relativity, and recency to the usage log data collected by each user session,
Based on the data obtained by the FRR model applying means for the analysis of interest, the interest of interest based on the individual behavior (including not only the product but also the object related to the user's use behavior) is extracted A display means for displaying interest from the highest ranking to the lowest ranking,
FRR model application method for analysis of preference field applying FRR model based on frequency, purchase, relativity, and recency to the interest ranking and score data collected by each user session,
The data obtained by the FRR model applying means for analyzing the preference field may be used as a preference field based on the personal behavior (the attribute information held by the object information related to the user's use behavior is not limited to the category or color of the product, A preference field indicating means for displaying preferences for preferences from a preference order to a preference order in order from the highest preference order to the lowest preference order display means.

Images