Cf Algorithm Based On User Characteristics In B2c Electronic Commerce Research And Applications Of Re

The size of the Internet users in China exceeded500million to513million and newInternet users throughout the year to55.8million，According to the report of CNNICJanuary29th China Internet Development Statistics,"As of the end of December2011.Internet penetration rate improved by four percentage over the last year, reaching38.3%.Analysis considered that the applications of e-commerce continues improved steadily，including online shopping, online payment, online banking, travel booking, E-commerceapplications continue to maintain a steady development trend in2011，the size of the onlineshopping users reach194million people, compared with the previous an increase of20.8percent at the end of the online payment users and online banking throughout the year alsoincreased by21.6%and19.2%of subscribers were167million and166million.The storage capacity of the IT system is far from less, not to mention the in-depthexcavation and analysis. However, in order to use these data is not an easy thing. Thispaper will recommend the system design as an example an integrated solution. In responseto data and real-time requirements of the major IT companies have proposed their ownsolution, many of whom are good systems and programming models, such as: googlehadoop technical framework, nosql databases and HDFS distributed storage technology,mapreduce programming mode, and the mahout. This article will be based on theframework of these emerging technologies to transform traditional era of real-timerecommendation engine to make it to meet large data requirements, to improve the hit rateof the recommended results.In this paper, improving the recommendation accuracy of the recommendation engine(RE) based on the the similarity user algorithm and the improved method of calculationsimilar to the user based on user characteristics. To give full consideration to the user’s ownsocial identity, the recommended data to improve individual needs. And to lucene platformspecific implementation, in this basis, the overall design of the chart to the recommendation engine. And to explain each part of the split. Finally, the technicaldifficulties in the specific implementation process and key points of the two practicalsolutions. Are based on the the redis+zookeeper’s main switching automatically from thebackup and the single point of failure "and" distribution technology based on the the dubbo+Lucene improve the index data automatically.