The Design And Implementation Of Personalized Search System Of Product

With the development of e-commerce, number of goods on shopping site is increasing. Using search engines to search product is becoming more and more difficult. Therefore, how to influence the search results according to the user’s own characteristics, their interests and their preferences, according to the user’s information and commodity information, allowing users to quickly find what they want is an urgent problem to be resolved. Based on the above background, this thesis designs a personalized search engine.Personalized Search includes query-based personalized search engine, personalized web weight based search engine and multi-personalized search engines. The method of user characteristics collection also have three options:server-side data-mining, users taking the initiative to provide and systems passive learning. Evaluating the advantages and disadvantages of these programs, combined with the company’s existing resources, This thesis selects the personalized web weight based search engine and server-side data-mining method. The system personalized three dimensions, namely, the gender of the user, the users’ price preferences and the user’s preferences on the product itself property. The three aspects at the same time affect commodity weight. Jobs needs to be done include the data-mining of the user gender, price preference, skin and clothing preferences, search keyword analysis, product attributes extraction, add a personalized index field and the calculation of the personality weighted. System use logistic regression model to predict user gender, use a combination of k-means clustering and Gaussian distribution model to classify price levels. Because of the large amount of data, the system is developed based on hadoop platform. Personalized feature weighting will add a personalized correction parameter, without changing other commodity weights.This thesis first introduces the development background of personalized search engine, and then introduced related technology in the achievement of the system. After analyzing system demand, we design a system’s overall architecture and each module of the system, and one by one features and design of each module. Finally, in accordance with the design of each module, give the detail implements of the system.Personalized search system based on the design and implementation of this article gives a more precise understanding of user intent, no doubt to facilitate the search for goods brought. It allows users to find what you want more efficiently, increase the user experience and make search more intelligent.