基于协同过滤的图书推荐系统

在商业领域,推荐系统被广泛用于向用户推荐符合其个人偏好的产品、服务或内容。借助这一技术建立图书推荐系统可以有效提高图书馆的服务水平。所提出的图书推荐系统是使用协同过滤技术通过对具有相似阅读习惯读者的借书数据进行偏好评分计算,从而为指定读者推荐符合其偏好的图书列表。为了解决推荐系统中所存在的数据稀疏性、评分的系统偏差以及图书偏好的量化等问题,该研究采用了矩阵分解、在评分中引入偏差值以及使用带时间戳的借阅记录生成偏好量化数值等解决方法。实验结果表明该推荐系统具有较好的准确度。

Book Recommendation System Based on Collaborative Filtering

The recommendation system is widely used in many commercial fields,it recommends products,services or content,to users according to his/her personal preference.Using this technology to establish a book recommendation system can effectively improve the service level of the library.The proposed book recommendation system uses collaborative filtering technology to perform preference score calculation on the borrowing data of students with similar reading habits,thereby recommending a set of books that meet the preferences of the designated students.In order to solve the problems of data sparsity,systematic deviation of ratings,and quantification of book preferences in the recommendation system,the study uses matrix decomposition,the introduction of deviation values in the ratings,and the use of time-stamped borrowing records to generate preference quantitative values.Experimental results show that the recommendation system has good accuracy.