item recommendation method based on preference sensitivity over purchase decision factor

摘要

Data processing using a computer, more specifically, a preferred product recommendation technology for recommending a product preferred by an individual by analyzing a purchase decision factor affecting a decision to purchase a product is disclosed. Among the purchasing decision factors, preference information having a sentence category is classified as a topic, and a topic tree is created using the association between the topics. Personalized preference sensitivity is calculated by reflecting the degree to which a specific customer's preference deviates from the average value of preference information for each purchase factor. A weight for each purchase decision factor may be calculated by preference sensitivity, and a preference score may be calculated for each product.