基于模糊聚类和改进混合蛙跳的协同过滤推荐*

由于传统的协同过滤推荐算法存在很多缺陷，如数据稀疏性、冷启动、低推荐精度等，提出了一种基于模糊聚类和改进混合蛙跳的协同过滤推荐算法。首先利用一种构造的基于时间的指数遗忘函数对原始评分数据进行处理；然后根据得到的基于时间衰退的评分矩阵对用户进行模糊C-均值(FCM)聚类，并找出与目标用户有较高相似性的前几个类作为候选邻居集；再用改进的混合蛙跳算法找到最近邻居集；最后求出目标用户对未参与项目的预测评分。经实验证明，该算法比其他一些算法的推荐精度要高，且由于数据稀疏性引起的不良影响也得到了有效的缓解。

Collaborative Filtering Recommendation based on Fuzzy Clustering and Improved Shuffled Frog Leaping algorithm

As the traditional collaborative filtering recommendation algorithm exists many defects, such as data sparseness, cold start and low recommendation accuracy, this paper proposes a collaborative filtering recommendation algorithm based on fuzzy clustering and improved Shuffled Frog Leaping algorithm. The algorithm first uses the constructed time-based exponential forgetting function to process the original score. Then, it cluster the users with fuzzy C-means (FCM) clustering according to the obtained scoring matrix based on time lag, and find the first few classes with higher similarity to the target user as candidate neighbor sets. And then use the improved Shuffled Frog Leaping algorithm to find the nearest neighbor sets. Finally, calculate the prediction score of the target user is not involved in the project. Experiments show that the proposed algorithm is more accurate than some other algorithms, and effectively alleviate the adverse effects due to data sparseness.