基于隐语义模型推荐算法的优化

Optimization of Recommendation Algorithm Based on Latent Factor Model

人们的生活已经离不开推荐系统了,而推荐算法的优劣则是推动推荐系统发展的重要因素。使用比较广泛的推荐技术有基于内容推荐、协同过滤以及混合推荐。但是以上推荐算法均存在精确率低,覆盖率窄等问题。论文融合了用户的情感因素以及物品的热门程度提出了一种基于潜在因子模型(LFM)的优化算法:基于动量的学习算法,最后通过实验证明改进后的算法比传统的算法在推荐精确度(Accuracy)以及覆盖率(Coverage)上都有明显的提升。

Recommendation system is indispensable to people’s life,and the pros and cons of recommendation algorithm is an important factor to promote the development of recommendation system. Among the more widely used recommendation techniques are association rules,collaborative filtering,and hybrid recommendations. However,the above recommendation algorithms all have problems such as low accuracy rate and narrow coverage. In this paper,an optimization algorithm based on latent factor model(LFM)is proposed by integrating the emotional factors of users and the popularity of items,which is momentum-based learning algorithm. Finally,experiments show that the improved algorithm has significant improvements in the Accuracy and Coverage compared with the traditional algorithm.