JP2003044750A - Internet shopping site for recognizing customer purchase pattern and performing optimum advertisement display - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that advertisement display which looks the same to anybody is performed because the advertisement display at an Internet shopping site grasps a customer group image in a fixed manner. SOLUTION: A support vector machine and a neural network are used to automatically learn a customer purchase pattern by using an 'information input (explanatory variable) by a question' and an 'information input (target variable) by purchase merchandise' until a previous time point in order to understand a customer group image, and optimum advertisement display for the 'information input by the question' at the present time point is performed.

Description

Description: BACKGROUND OF THE INVENTION [0001] 1. Field of the Invention [0002] The present invention relates to an advertisement display on an Internet shopping site. 2. Description of the Related Art As a conventional purchasing support method for Internet shopping, as shown in FIG. 8, the same advertisement is displayed regardless of who sees it. [0003] However, according to the above-mentioned prior art, the group image of the customer is constantly captured, and the same advertisement is displayed by anyone. However, it is an object of this technology to provide an advertisement display desired by a customer group image. [0004] In order to solve the above-mentioned problems, the invention according to claim 1 is for understanding a group image of customers.
Perform "information input by question" as shown in FIG.
This information at the present time is stored in a database with all the information input by question (explanatory variables) and the information input by purchased products (objective variables) up to the previous time, and learned with a support vector machine, neural network, etc. Internet shopping site that inputs to the learned support vector machine or neural network and displays advertisements of "products predicted to be purchased" output by the learned support vector machine or neural network to support purchasing . FIG. 1 shows an embodiment of the present invention. Since the learning unit requires a certain number of samples, learning cannot be performed in the initial state. Therefore, the advertisement section is emptied in the initial state, or as many learning samples as possible, learning samples are prepared based on a survey such as a questionnaire. [0006] When the sample is empty, only information is stored and stored in a database until learning samples are sufficiently collected. When sufficient learning samples are collected, learning is performed using a neural network or a support vector machine. When a learning sample is prepared, a learning sample is prepared, stored in a database, and learning is performed. [0008] The database used here is as shown in Fig. 5. If there is no "information input based on purchased merchandise", it means that the customer did not purchase, and it is used for learning of neural networks and support vector machines. Absent. This database is used in a learning unit using a neural network or a support vector machine. When storing information in the database, as shown in FIG. 6, "information input by question" is temporarily stored in the personal computer of the client, and at the same time as information of "information input by purchased product" at the time of purchase. The “information input by question” which is temporarily stored is transmitted and stored in the database. [0010] "Information input by question" is input to a learned neural network or a support vector machine, and an advertisement of a product predicted to be purchased by the customer is displayed. "Effects of Embodiment" According to this embodiment, unlike a general advertisement display, an optimum advertisement display for a customer can be realized. "Other Embodiments" Although the product is a product in FIG. 3, a content or a character may be used. Although the objective variable is “information input by purchased product”, a variable such as a vote or a score may be used. As described above, according to the present invention, since a neural network and a support vector machine are used, an optimum advertisement display can be performed.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing an embodiment of the present invention. FIG. 2 is an input screen on a Web of a question section. FIG. 3 is a screen on a Web of a display section. Screen of purchase department (detailed display) on the Web [Fig. 5] Table of input example of database [Fig. 6] Table of flow until storage of information in database [Fig. 7] Flow of learning update of learning unit FIG. 8 is a diagram showing an example of a conventional advertisement on a Web page. [Description of Signs] 1 A part where a question is asked is as shown in FIG. 2 A part where learning is performed using a neural network or a support vector machine 3 An optimal advertisement and a product configuration screen predicted using a learning unit are displayed In the part where you do something like Fig. 3 4 If you want to display the details of the product and purchase it, click on any button and the part like Fig. 4 5 "Input information by question" and "Depend on the product actually purchased Where to store "input information" 6 Screen to select by clicking the check box on the questionnaire input screen 7 Button to send "input of information by question" and click to jump to the product configuration screen 8 Learning unit 9 Displayed advertisement using the learning unit on the product configuration screen 10 Displayed on the product detailed information display screen Detailed information display 11 Button clicked at the time of purchase 12 "Information input by question" stored in database 13 "Information input by purchased product" stored in database 14 Internet terminal used by customer 15 Information stored Internet server 16 that displays the same advertisement no matter who sees it on the conventional advertisement on the Web 17 This is a product configuration screen, and when each product is clicked, a detailed screen is displayed

Claims (1)

Claims 1. In order to understand a group image of customers, "information input by question" is performed, and the information at the present time is converted into all "information input by question (explanatory variable)" up to the previous time. ”And“ Information input (purchased variables) based on purchased products ”are stored in a database and trained using a support vector machine or neural network, and then input into a trained support vector machine or neural network to provide the learned support. An Internet shopping site that supports purchasing by displaying advertisements of "products predicted to be purchased" output by vector machines and neural networks.