基于社区划分的现代文学作品个性化推荐算法

A Personalized Recommendation Algorithm of Modern Literature Works Based on Community Division

文学作品作为人类文明的重要表现形式之一,在互联网技术的推动下公众接受文学作品的方式越来越依赖平台的推荐。为了提高现代文学作品推荐算法的准确性,提出了加权信息增益算法(DWIG)与选择算法(TF-IDF-DW),其中加权信息增益算法是对特征项在类间和类内的分布特征优化,选择算法是对特征项位置分布权重的优化。在传统推荐算法的基础上,结合读者评论文本分析情况,按照社区的方式对读者进行划分。根据社区划分与读者评论文本的处理结果,对读者的阅读喜好进行综合预测评分,从而提升推荐的准确性。最后的实验结果显示,传统协同过滤算法的MAE值为1.8,而此次研究所提出的改进算法的MAE值只有0.5,该结果表明此次研究提出的推荐算法极大地提高了智能推荐系统的准确度。

As one of the important forms of human civilization,literary works are accepted by the public more and more depending on the recommendation of the platform under the promotion of the Internet technology.In order to improve the accuracy of the recommendation algorithm for modern literary works,a weighted information gain algorithm(DWIG)and a selection algorithm(TF-IDF-DW)are proposed.Among them,the weighted information gain algorithm optimizes the distribution characteristics of feature items among classes,and the selection algorithm optimizes the distribution weights of feature items.Based on the traditional recommendation algorithm,combined with the analysis of readers'comments,readers are divided according to the way of community.According to the results of community division and reader comment text processing,readers'reading preferences are comprehensively predicted and scored,so as to improve the accuracy of recommendation.The final experimental results show that the MAE value of the traditional collaborative filtering algorithm is 1.8,while the MAE value of the improved algorithm proposed in this study is 0.5,which shows that the proposed recommendation algorithm greatly improves the accuracy of the intelligent recommendation system.