英语翻译推荐技术作为一种信息过滤的重要手段,是当前解决信息超载问题的非常有潜力的方法.推荐技术最典型的应用是在电子商务领域.推荐技术可根据用户的兴趣、爱好推荐顾客可能感兴

英语翻译推荐技术作为一种信息过滤的重要手段,是当前解决信息超载问题的非常有潜力的方法.推荐技术最典型的应用是在电子商务领域.推荐技术可根据用户的兴趣、爱好推荐顾客可能感兴
英语翻译
推荐技术作为一种信息过滤的重要手段,是当前解决信息超载问题的非常有潜力的方法.推荐技术最典型的应用是在电子商务领域.推荐技术可根据用户的兴趣、爱好推荐顾客可能感兴趣或满意的商品(如书籍、音像等).挖掘用户的潜在需求转化为现实需求,从而达到提高产品销售量的目的.个性化推荐系统是建立在海量数据挖掘基础上的一种高级商务智能平台,以帮助电子商务网站为其顾客购物提供完全个性化的决策支持和信息服务.因此,研究推荐技术意义重大.
本文首先分析了当前推荐算法的现状,提出了研究本课题的意义,然后介绍了推荐系统的研究现状及其存在问题,对现有的推荐系统进行较为详细的梗概.在梗概的基础上,引出对协同过滤的详细综述.并就其研究现状及存在问题、相似性比较方法等进行阐述.接着着重研究了不确定近邻和基于项目评分预测协同过滤算法,给出了详细的算法思想和公式.最后,是不确定近邻和基于项目评分预测协同过滤算法的实现：用matlab语言分别编写出两种算法,并且在matlab上运行.
实验结果表明,在本文所采用的数据集条件下,不确定邻居的方法预测准确率要高于基于项目评分的,不确定近邻的方法随着参与者的增多预测准确率有所提高,当用户超过某一值时达到稳定值.
关键词：推荐算法；协同过滤；相似度：不确定近邻；项目评分

英语翻译推荐技术作为一种信息过滤的重要手段,是当前解决信息超载问题的非常有潜力的方法.推荐技术最典型的应用是在电子商务领域.推荐技术可根据用户的兴趣、爱好推荐顾客可能感兴
Recommendation Algorithm, as one of the most important methods of Information Filtering (IF), is considered for having a very high potential in resolving the current "information overload" problem. One of the most prominent applications of Recommendation Algorithm is in the Electric Commerce (E-Commerce) field. The Recommendation Algorithm method can suggest possible products that may be of interest to a customer (such as books and CDs) based on the customer's interest and preferences. It identifies the potential needs of the users, and turn them into actual needs, in order to reach the targets of improving the sales volume of the product. The personalized Recommender System is a high-end Business Intelligence platform that is based on a Mass Data Mining approach, in order to help E-Commerce sites in supporting completely customized customer shopping experience and decision-making, as well as providing information. This suggests the high significance of developing the Recommendation Algorithm.
This report starts with introduces the state quo of Recommendation Algorithm, suggests the importance of studying the subject. It also examines the development of Recommendation Algorithm and the existing problems, and outlines the similiarity comparasion methods. Based on the outlines, it provides a comprehensive overview of Collaborative Filtering and studies the existing conditions and issues. It then focuses on the examination of the Collaborative Filtering Recommendation Algorithm based on Uncertain Neighbors and Item Rating Prediction, and presents the detailed methodology in support of the Recommendation Algorithm and its formulas. At last, as realized in the Collaborative Filtering Algorithm based on Uncertain Neighbors and Item Rating Prediction, two algorithms are developed using MATLAB language to be implemented in MATLAB.
The results of the study unveils that, under the conditions of the data collected by this report, the accuracy rate of the Uncertain Neighbors approach is higher than the Item Rating Prediction approach. Also, when the Uncertain Neighbors approach is being applied, the accuracy rate becomes higher as the number of the participants in the study increases; and the accuracy rate becomes constant as the number of participants reaches to a certain value.
全是自己人工翻译的.