基于协同过滤的美食推荐算法

为了解决传统的基于用户的协同过滤算法中的数据稀疏性问题，提高推荐的准确率，本文对推荐算法进行了改进并将改进后的算法应用在美食推荐领域。首先，利用均值中心化方法对实验数据进行处理，减少因个人评分习惯差异造成的推荐误差。然后，通过使用改进的空值填补法降低评分矩阵的稀疏性。最后，在计算相似度时引入了遗忘函数和用户间的信任度，进一步提高了推荐系统的准确性。实验表明，本文提出的改进算法比传统算法有更高的准确率，并得出了在推荐过程中考虑用户和项目外的其他因素以及针对不同的数据信息采用不同的算法，都有利于提高推荐准确率的重要结论。

Food Recommendation Algorithm Based On Collaborative Filtering

In view of the problem of data sparseness in the traditional user-based collaborative filtering algorithm, improved the recommended accuracy, this paper put forward an improved algorithm and used this algorithm to the field of food recommendation. Firstly, in order to reduce the recommended error caused by different personal rating habits, this paper used the mean centralized method to dispose the score data. Secondly, with the improved null values fill method to reduce the sparse of the rank matrix. Finally, when calculating the similarity between users, this paper considered the factors of the forgotten function and trust relationship between users, in order to improve the accuracy of recommendation system. The experiment showed that the proposed algorithm can get higher accuracy than the traditional algorithm, and it is concluded that in the process of recommended, consider about other factors expect users and items as well as use different algorithms for different data information are beneficial to improve the accuracy of recommendation.