Abstract: According to data sparsity and recommendation accuracy problem in recommender systems, a matrix decomposition algorithm is proposed to preserve the original data structure. Using the projection structure of Non-Negative matrix factorization algorithm, combining the trust propagation model and the personal compact calculation model and user rating prediction model, a fusion of influential individuals projection of the structure of non-negative matrix factorization recommendation algorithm is proposed to improve the precision of the recommendation system. Through the experiments prove that, compared with other algorithms, this algorithm not only reduces the MAE and RMSE, but also improves the prediction accuracy of the system.

Key words: Personal Compact, recommender systems, trust propagation, non-negative matrix factorization

摘要： 针对推荐系统中存在的数据稀疏性和推荐准确性问题，利用信任传递思想，融合个体影响力计算模型和用户评分预测模型，使用结构投影非负矩阵分解推荐算法，采用随机梯度下降逼近方法，提出了一种以保留原始数据结构特征为目的、融合个体影响力和信任传递的结构投影非负矩阵分解推荐算法TP-SPNMF。通过多组对比实验证明，相比其他算法，TP-SPNMF算法不仅降低了MAE和RMSE，还提高了系统的预测准确性。

关键词: 个体影响力, 推荐系统, 信任传递, 非负矩阵分解