As well as the program spread prediction system, method, and program impact estimation system, method and

摘要

PPROBLEM TO BE SOLVED: To provide a circulation prediction system capable of predicting a time when a customer starts to use an article or service for each customer. PSOLUTION: A customer DB 2 stores a set of individual data for each customer including a use starting period of the article etc. or a fact that the article etc. is not yet used, and items indicating attributes of customers. A test data generation part 51 generates test data including individual data about customers who do not start use at a present time using the customer DB 2. A learning data generation part 52 attaches a first label to the individual data about the customers who start the use of the article etc. at the present time, attaches a second label to the individual data about the customers who do not start the use of the article etc. at the present time, generates learning data, and a classifier generation part 53 generates a classifier from the learning data. A test data label determination part 54 determines labels to each piece of individual data by collating the test data with the classifier. PCOPYRIGHT: (C)2010,JPO&INPIT