针对云平台协同推荐的近邻项目最优临界点优化

Optimal Critical Point Optimization of Nearest Neighbor Projects Recommended for Cloud Platform Collaboration

针对单一的基于用户或者基于项目的推荐算法在个性化推荐的应用中还存在精度不高、推荐结果不佳的问题,本文提出了一种近邻项目最优临界点优化的云平台协同推荐模型。首先根据所有的使用客户本身存在一定的偏向爱好,计算项目之间的相似度,得到用户对物品的相似度评价,然后采用线性回归的方法对相似度评价结果进行重新预测估计,最后对多个近邻用户和多个近邻项目的最优临界点进行优化,以提高推荐精度。仿真实验结果表明,本文提出的改进模型在个性化推荐的应用中,具有更高的稳定性和推荐精度。

For a single user-based or project-based recommendation algorithm, there is still a problem of poor accuracy and poor recommendation in the application of personalized recommendation. This paper proposes a cloud platform collaborative recommendation model with optimal critical point optimization for neighborhood projects. First of all, according to all the use of the customer itself there is a certain degree of preference, calculate the similarity between the items, the user's similarity evaluation of the items, and then use the linear regression method to re-predict the results of similarity estimates,Neighbor users and multiple neighboring projects to optimize the optimal threshold to improve the recommended accuracy. The simulation results show that the improved model proposed in this paper has higher stability and recommended precision in the application of personalized recommendation.