面向个性化产品服务方案的推荐方法与应用

Recommendation Method and Applications for Personalized Product Service Plans

针对传统隐语义模型(LFM)未考虑数据库动态更新从而影响方案推荐结果的问题,提出动态更新机制的加权LFM用于推荐个性化产品服务方案。针对余弦相似度计算忽略个体差异的问题,提出采用云滴距离测度与云的余弦相似度加权后的综合相似度,预测并填充空缺数据,减少数据稀疏性;采用加权LFM推荐产品服务方案,以约束新用户兴趣差异性,提高推荐精度;采用差值平均法更新推荐结果。

Traditional LFM didn’t consider the database dynamic update which would affect the program recommendation results,so a weighted LFM was proposed based on the dynamic update mechanism to recommend the personalized product service plans.Firstly,aiming at the problems of cosine similarity neglecting individual differences,a comprehensive similarity weighted by cosine similarity between cloud drop distance measure and cloud was proposed to predict and fill vacancy data and reduce data sparsity.Then,the difference of interests of new users could be restrained and the recommendation accuracy could be improved by using the weighted LFM recommendation service scheme.Finally,the proposed method was updated with the method of difference averaging,and the efficiency of different recommended algorithms was compared with mean absolute error(MAE).The practicality and validity of the proposed algorithm were verified by examples.