International Business Machines Corporation
Scenario enhanced search with product features

Abstract:

A system is provided that includes a server having a first engine, a second engine, a search word warehouse, and a purchasing engine. The first engine (i) maps topics to product features of products using a topic hotness method which computes a hotness degree of a product on a topic from (a) product identifiers, (b) product features, (c) customer comments, and (d) topics with relevant key words, and (ii) builds a model providing a relevance degree of a topic to a product feature. The model is built based on the hotness degree. The second engine extracts remarkable topics based on product features using a method applied to the relevance degree of the topic to the product feature to identify remarkable topics for each product. The search word warehouse stores the remarkable topics for each product. The purchasing engine conducts purchases of products identified in searches using the remarkable topics.