基于多兴趣度的图书借阅推荐系统研究与设计

针对数字图书馆图书借阅推荐系统中通常采用的协同过滤算法中存在的不足,分析了该算法存在在用户多兴趣情况下推荐时难免会产生偏差,从而降低预测推荐准确率问题,进而提出了一种基于用户多兴趣度的图书借阅推荐系统推荐算法的设计思路。该算法首先对待推荐图书进行分类,然后根据用户的借阅情况计算其对每类图书的兴趣度,将对每类图书兴趣度相似的用户聚为一类,在小范围内寻找最近邻居,最后根据用户对不同类别书目的兴趣度作为权重值来产生对该用户的每类图书的推荐数目。该算法不仅能够解决用户多兴趣度问题,同时也可减少计算工作量,提高推荐的准确率。

Books based on more interested in degree of research and design recommendation system

This recommendation for the digital library book lending system,collaborative filtering algorithms commonly used in the shortcomings of the existence of the algorithm is more interested in the case the user will inevitably bias when recommended,thus reducing the problem of prediction accuracy of recommendation,and then made more interested in a degree based on user recommendation system to borrow books recommended algorithm design ideas.firstly,classification of books recommended treatment,then the situation according to the user to calculate the effect of borrowing books of interest for each degree level will be similar to each class of users interested in books together as a class,in a small area to find the nearest neighbor,finally users interested in different types of bibliographic degree as to produce the weights for each user the number of recommended books.The algorithm can solve the problem of the user more degrees of interest,but also can reduce the computational workload and improve the recommendation accuracy.