基于改进相似性度量的项目协同过滤推荐算法

发布时间：2018-07-23 12:51

【摘要】：针对传统协同过滤推荐算法遇到冷启动情况效果不佳的问题,提出一种基于项目相似性度量方法(IPSS)的项目协同过滤推荐算法(ICF_IPSS),其核心是一种新的项目相似性度量方法,该方法由评分相似性和结构相似性两部分构成:评分相似性部分充分考虑两个项目评分之间的评分差、项目评分与评分中值之差,以及项目评分与其他评分平均值之差;结构相似性部分定义了共同评分项目占所有项目比重,并惩罚活跃用户的逆项目频率(IIF)系数。在Movie Lens和Jester数据集下测试算法准确率。在Movie Lens数据集下,当近邻数量为10时,ICF_IPSS的平均绝对偏差(MAE)和均方根误差(RMSE)分别比基于Jaccard系数的均方差异系数的项目协同过滤算法(ICF_JMSD)低3.06%和1.20%;当推荐项目数量为10时,ICF_IPSS的准确率和召回率分别比ICF_JMSD提升67.79%和67.86%。实验结果表明,基于IPSS的项目协同过滤算法在预测准确率和分类准确率方面均优于基于传统相似性度量的项目协同过滤算法,如ICF_JMSD等。
[Abstract]:In order to solve the problem that the traditional collaborative filtering recommendation algorithm has a poor effect on cold start, a project collaborative filtering recommendation algorithm (ICF_IPSS) based on item similarity measurement method (IPSS) is proposed, the core of which is a new project similarity measurement method. The method consists of two parts: score similarity and structural similarity: the score similarity part fully considers the difference between the two items, the difference between the item score and the median score, and the difference between the item score and other average scores; The structural similarity part defines the proportion of common scoring items to all items and penalizes the inverse item frequency (IIF) coefficient of active users. The accuracy of the algorithm is tested under Movie Lens and Jester data sets. Under the Movie Lens dataset, When the number of nearest neighbors is 10:00, the average absolute deviation (MAE) and root mean square error (RMSE) are 3.06% and 1.20% lower than those of the ICF_JMSD algorithm based on Jaccard coefficient, respectively, and the accuracy and call rate of ICFIPSS are lower when the number of recommended items is 10:00. The return rate was 67.79% and 67.86% higher than that of ICF_JMSD, respectively. The experimental results show that the project collaborative filtering algorithm based on IPSS is superior to the item cooperative filtering algorithm based on traditional similarity measure in prediction accuracy and classification accuracy, such as ICF_JMSD, etc.
【作者单位】：东北农业大学工程学院;东北农业大学理学院;
【基金】：公益性行业(农业)科研专项二级任务(201503116-04-06) 黑龙江省博士后基金资助项目(LBH-Z15020) 国家科技支撑计划专题任务(2014BAD12B01-1-3) 农业部农业水资源高效利用重点实验室开放基金资助项目(2015004) 黑龙江省哲学社会科学研究规划年度项目(16YB17)~~
【分类号】：TP391.3

【相似文献】