E-Commerce Web Database Approximate Query Results Ranking

In search of e-commerce network database, users often have vague or imprecise ideas, formulation of queries can not accurately convey the intent of their query, the query to get the relevant answers to relax, to meet user needs and preferences, has become an important method. However, inquiries to relax, the user may be faced with too many answers. In response to queries from the approximate answer e-commerce Web database returns too many problems, a new method for approximate query results sorting. Based on database query history, we hypothesized that users care about each property, and the corresponding weight assigned to it. Then, depending on the property value of similarity between a query based on the results of approximate satisfaction ranking method. Secondly, satisfaction with the same tuple, each attribute value is not specified, the user\´s "satisfaction" is given a score. Proved that the method to obtain satisfactory results. Finally, preliminary experiments show its efficiency and the quality of the ranking algorithm.