Implementation for full-text search and rank optimization in e-commerce

CNBAB is a Chinese e-commerce platform. CNBAB trust model aims to help agent build a credible trading environment. In CNBAB, there is a method for evaluating initial trust value of direct trust and recommender trust and updating them according to the outcome of interaction; there are a recommend protocol and algorithm for finding multi trust paths and a method for computing indirect trust value from trust paths. A good resource search engine can help users find resources quickly. Due to the limitations of resource search based on relational database query, we build a full-text resource search engine based on Coreseek and mmseg in CNBAB. We adopt two methods to sort search results. One approach regards structured information of resources and adjusts influencing factors so that the final ranking results reflect the logical relationship between resources, but has no optimization for the same sort of resources. Another approach uses direct trust value to deal with ranking for the same sort of resources. Combined with the two approaches, search results are stable, ranking optimization effect is obvious.