Research And Practice Of Diversity In Personalized Recommendation System

The 21 st century is the era of information technology,with the development of computer technology and modern e-commerce,e-tailers come into people's lives with it's diversify products and fast delivery features.Internet retailers has long-tail effect,that is 80% of corporate sales come from 20% of the goods,and if possible to improve the long tail of retailers selling goods,the turnover of online retailers will grow exponentially,and can unique meet user preferences.In this context,the research for the retail sales volume of long tail items and excavation unique user preferences has profound significance,therefore,how to make personalized recommendation system users more comprehensive and recommender all items in the system and improve personalized recommendation system's diversity is a challenging research topic.In this paper,against the diversity problem in personalized recommendation system,we put forward the corresponding solutions.The main work of this paper is as follows:(1)For personalized recommendation system follow recommendation accuracy and ignoring the diversity issue,we put forward a personalized recommendation system diversity solution design.By adding a control threshold sort method after the general personalized recommendation system's core recommendation module,and based on the common core personalized recommendation,using a threshold value control method and sorting technology reconstruct its recommended list,the method can reduce less recommendation accuracy while substantial increase in overall diversity recommendation result.This scheme has minimal impact of the original personalized recommendation system,it's a more flexible and convenient method of expansion,capable of binding with different personalized recommendation algorithm design and able to adapt to different personalized recommendation system.(2)For a personalized recommendation system in the long tail effect problems,studying the factors of long tail items associated with and proposed long tail items sorting method.In the network retailers,taking into account the emergence of the long tail items is a common result of many factors,by researching the reason long tail items appear,characterize the long tail items appear several factors,and to quantify these factors,use clear value to measure these factors,make it based on the control threshold of personalized recommendation system sorting method to sort the list basis.The base was added to a sorted list based on the control threshold of personalized recommendation system sorting method,the experiment proved that this method can help to solve the personalized recommendation system in the long tail of the problem.(3)Design and implement an offline Douban movie personalized recommendation system.In the system environment,the completion of the work of watercress personalized recommendation data and verify the correctness and effectiveness of the proposed method.