基于特征嵌入的去流行度偏差混合推荐算法

Hybrid recommendation algorithm for removing popularity bias based on feature embedding

针对数据不均衡条件下贝叶斯个性化排序算法生成的推荐列表中存在强流行度偏差的问题,提出基于特征嵌入的去流行度偏差混合推荐算法。首先,利用卷积神经网络提取用户、物品特征确定用户偏好,并依据用户偏好对原始不均衡数据进行评分填充;其次,将卷积神经网络提取的用户偏好特征嵌入到贝叶斯个性化排序算法中进行混合推荐;最后,用评分填充数据训练混合推荐模型,得到去流行度偏差的个性化排序列表。为了验证算法的性能,在公开数据集MovieLens-100K和MovieLens-1M上进行分析与对比实验,实验结果显示流行度偏差降低了约50%~60%,精确度提高了约一倍。

In order to solve the problem of strong popularity bias in recommendation lists generated by Bayesian personalized ranking(BPR) algorithm under the condition of unbalanced data,this paper designed a hybrid recommendation algorithm based on feature embedding to remove popularity bias.Firstly,this paper used convolutional neural network to extract user and item features to determine user preferences,and filled the original unbalanced data according to user preferences.Secondly,it embedded the user preference features extracted from convolutional neural network into the BPR algorithm for hybrid recommendation.Finally,it trained the mixed recommendation model with score filled data,and obtained the personalized ranking list without popularity bias.In order to verify the performance of the algorithm,this paper conducted the analysis and comparison experiments on MovieLens-100 K and MovieLens-1 M.Experimental results show that the popularity bias is reduced by about 50%~60% and the accuracy is improved by about twice.