互联网拥有全世界最大的信息资源，随着互联网信息的飞速发展，网络信息呈爆炸式递增趋势。丰富的资源给人们的学习和生活带来诸多的便利，但同时也引发信息过载等问题。对于缺乏明确目标的用户，如何在短时间内找到满意的信息资源，成为当前研究的热点。协同过滤是推荐系统中使用最为广泛的一种算法，其中基于用户的协同过滤算法的思想是通过计算用户之间的相似度，找到目标用户的相似邻居集合，通过分析相似用户对某些商品的评分数据，预测目标用户对未评分项目的分值，选取评分最高的若干项目进行推荐。上述算法在使用过程中存在相似度计算不准确和评分数据稀疏等问题，因此需对相似度计算进行优化改进，主要的工作内容如下：

针对协同过滤算法中自身存在相似度计算不准确，不能较好的反映出用户之间的关联关系，导致推荐结果出现偏差，提出一种融合相似度的协同过滤推荐算法。该方法在传统相似度的基础上，通过结合个体用户公共项目占比、整体用户公共项目占比和用户公共项目评分偏差构建用户相似度模型；借助用户对项目的评分习惯构建用户偏好相似度模型，同时引入热门商品惩罚因子对用户偏好相似度计算公式进行优化；最后将用户相似度模型和用户偏好相似度模型加权融合得到新的相似度模型，采用新的相似度模型预测评分并向目标用户进行推荐。实验结果表明，新的相似度模型可以提升用户相似度计算的准确性，能弥补传统相似度计算方法存在的弊端和不足。   

针对用户评分数据稀疏问题对推荐结果造成的不良影响，提出一种融合Slope One和权威用户相似度的协同过滤推荐算法。首先提出一种基于用户的Slope One算法；然后通过对现实生活中的用户关系进行分析，分别从专业度、知名度和评分信任度三个方面来引入权威用户，将权威用户和Pearson相似度相结合得到权威用户相似度，并将权威用户相似度作为权重和基于用户的Slope One算法进行加权融合去预测评分；最后将预测评分构建成新的评分矩阵，在新的评分矩阵上使用优化后的Jaccard相似度去寻找最近邻居集合并完成项目推荐。实验结果表明，改进算法有助于提升系统的推荐性能，可有效的缓解数据稀疏性问题。

The Internet has the world's largest information resources, with the rapid development of Internet Information, network information is an explosive increasing trend. Abundant resources bring a lot of convenience to people's study and life, but at the same time it also leads to information overload and other problems. How to find satisfactory information resources in a short period of time has become a hot spot of current research for the users who lack clear goals. Collaborative filtering is one of the most widely used algorithms in recommender systems, in which the idea of user-based collaborative filtering algorithm is to find the similar neighbor set of target users by calculating the similarity between users, by analyzing the rating data of similar users on some commodities, the paper predicts the scores of unrated items by the target users, and selects some items with the highest scores for recommendation. The above algorithm has some problems such as inaccurate calculation of similarity and sparse score data, so it is necessary to optimize the calculation of similarity. The main work contents are as follows:

Aiming at the inaccuracy of similarity calculation in collaborative filtering algorithm, which can not reflect the association relationship between users well and lead to the deviation of recommendation results, a collaborative filtering recommendation algorithm combining similarity was proposed. Based on the traditional similarity, the user similarity model was constructed by combining the proportion of individual users' public projects, the proportion of overall users' public projects and the deviation of user's public projects. The user preference similarity model was built based on the user's scoring habits of the project, and the penalty factor of popular goods was introduced to optimize the calculation formula of user preference similarity. Finally, the user similarity model and the user preference similarity model are weighted and fused to get a new similarity model, which is used to predict the score and make recommendations to target users. The experimental results show that the new similarity model can improve the accuracy of user similarity calculation and make up for the drawbacks and deficiencies of traditional similarity calculation methods.

 (2) Aiming at the negative impact of sparse user score data on recommendation results, a collaborative filtering recommendation algorithm is proposed which integrates the similarity between Slope One and authoritative users. Firstly, a user-based Slope One algorithm is proposed. Then, by analyzing the relationship between users in real life, authoritative users are introduced from three aspects: professionalism, popularity and rating trust. Authoritative users are combined with Pearson's similarity to obtain authoritative users' similarity. The authoritative users' similarity is used as the weight to predict the score by weighted fusion with user-based Slope One algorithm. Finally, the predicted score is constructed into a new scoring matrix, on which the optimized Jaccard similarity is used to find the nearest neighbor set and complete the project recommendation. The experimental results show that the improved algorithm is helpful to improve the recommendation performance of the system and can effectively alleviate the data sparsity problem.

[1]CNNIC发布第49次《中国互联网络发展状况统计报告》[J]. 新闻潮, 2022(02): 3.

[2]郭佳, 黄程松. 国外网络环境中信息过载研究进展[J]. 情报科学, 2018, 36(07): 170-176.

[3]Wang T, Shi X, Chen Q. Research on the Vocabulary Relevancy Algorithm of the Improved Search Engine[J]. Journal of Physics: Conference Series, 2020, 1682(1): 1270-1279.

[4]于蒙, 何文涛, 周绪川, 等. 推荐系统综述[J]. 计算机应用, 2022, 42(06): 1898-1913.

[5]Liu H, He J, Wang T, et al. Combining user preferences and user opinions for accurate recommendation[J]. Electronic Commerce Research and Applications, 2013,12(1):14-23.

[6]Vijayakumar V, Subramaniyaswamy V, Logesh R.et al. Effective Knowledge Based Recommender System for Tailored Multiple Point of Interest Recommendation[J]. International Journal of Web Portals (IJWP), 2019, 11(1): 1-18.

[7]邓爱林, 朱扬勇, 施伯乐. 基于项目评分预测的协同过滤推荐算法[J]. 软件学报, 2003(09): 1621-1628.

[8]陈碧毅, 黄玲, 王昌栋, 等. 融合显式反馈与隐式反馈的协同过滤推荐算法[J]. 软件学报, 2020, 31(03): 794-805.

[9]赵俊逸, 庄福振, 敖翔, 等. 协同过滤推荐系统综述[J]. 信息安全学报, 2021, 6(05): 17-34..

[10]Goldberg D, Nichols D, Oki B M, et al. Using collaborative filtering to weave an information tapestry[J]. Communications of the ACM, 1992, 35 (12): 61-70.

[11]Konstan J A, Miller B N, Maltz D. et al. GroupLens:Applying Collaborative Filtering to Usenet News[J]. Commun. ACM, 1997, 40(3): 77-87.

[12]Shahbi C, Banaei-Kashani F, Chen Y S, et al. Yoda:an accurate and scalable web-based recommendation system[C]. In:Proceedings of the 6th International Conference on Cooperative Information Systems. London, 2001: 418-432.

[13]Chen H, Li Z, Hu W. An improved collaborative recommendation algorithm based on optimized user similarity[J]. The Journal of Supercomputing, 2016, 72(7): 2565- 2578.

[14]Koohi H, Kiani K. A new method to find neighbor users that imporves the performance of collaborative filtering[J]. Expert Syst Appl. 2017, 83: 30-39.

[15]Singh PK, Sinha M, Das S, et al. Enhancing recommendation accuracy of item-based collaborative filtering using Bhattacharyya coefficient and most similar item[J]. Applied Intelligence, 2020, 50(12): 4708-4731.

[16]赵厉宇哲, 刘学军, 徐新艳. 融入专业度和用户相似性的跨域推荐算法[J]．计算机工程与设计, 2019, 40(01): 136-142.

[17]Zhou W, Li R, Liu W. Collaborative filtering recommendation algorithm based on improved similarity[C]//IEEE 5th Information Technology and Mechatronics Engineering Conference(ITOEC), Chongqing, China, 2020: 321-324.

[18]韩胜宝, 伊华伟, 李晓会, 等. 基于融合相似度和层次聚类的冷启动推荐算法[J].小型微型计算机系.

[19]张瑞典, 钱晓东. 用余弦相似度修正评分的协同过滤推荐算法[J]. 计算机工程与科学. 2020, 42(06): 1096-1105.

[20]张凯辉, 周志平, 赵卫东. 结合CFDP与时间因子的协同过滤推荐算法[J]. 计算机工程与应用, 2020, 56(15): 80-85.

[21]田震, 潘腊梅, 尹朴, 等.深度矩阵分解推荐算法[J]. 软件学报, 2021, 32 (12): 3917-3928.DOI:10.13328/j.cnki.jos.006141.

[22]Davtalab M, Alesheikh A A. A POI recommendation approach integrating social spatio-temporal information into probabilistic matrix factorization[J]. Knowledge and Information Systems, 2021, 61(1): 65-85.

[23]Liu H Z, Wang W, Zhang Y H, et al. Neural Matrix Factorization Recommendation for User Preference Prediction Based on Explicit and Implicit Feedback[J]. Computational Intelligence and Neuroscience, 2022, 2022.

[24]Najafabadi M K, Mohamed A, Nair M A, et al. An Effective Collaborative User Model Using Hybrid Clustering Recommendation Methods[J]. ISI, 2021, 26(2).

[25]王菲, 黄刚, 朱峥宇. 基于信任聚类的协同过滤推荐算法[J]. 计算机技术与发展, 2019, 29(05): 22-26.

[26]苏庆, 章静芳, 林正鑫, 等. 改进模糊划分聚类的协同过滤推荐算法[J]. 计算机工程与应用, 2019, 55(05): 118-123.

[27]韩亚楠, 曹菡, 刘亮亮. 基于评分矩阵填充与用户兴趣的协同过滤推荐算法[J]. 计算机工程, 2016, 42(01): 36-40.

[28]王志远, 王兴芬. 基于用户兴趣差异改进矩阵填充的个性化推荐算法[J]. 计算机应用与软件, 2020, 37(12): 224-230+237.

[29]沈学利, 李子健, 赫辰皓.基于评分填充与信任信息的混合推荐算法[J]. 计算机应用, 2020, 40(10): 2789-2794.

[30]任永功, 王思雨, 张志鹏.缓解数据稀疏问题的协同过滤混合填充算法[J]. 模式识别与人工智能, 2020, 33(02): 166-175.

[31]周寅莹, 章梦怡, 余敦辉, 等.融合信任隐含相似度与评分相似度的社会化推荐[J/OL].计算机应用.http://kns.cnk i.net/kcms/detail/51.1307.TP.20211231.1729.018. html.

[32]Bathla G, Aggarwal H, Rani R. A graph-based model to improve social trust and influence for social recommendation[J]. The Journal of Supercomputing: An International Journal of High-Performance Computer Design, Analysis, and Use, 2020, 76(9).

[33]Ahmadian S, Meghdadi M, Afsharchi M. A social recommendation method based on an adaptive neighbor selection mechanism[J]. Information Processing and Management, 2018, 54(4).

[34]杨鹏, 邵堃, 霍星, 等. 融合用户隐含偏好的社会化推荐算法[J]. 小型微型计算机系统, 2019, 40(10): 2039-2045.

[35]刘超, 赵文静, 贾毓臻, 等.基于改进的BiasSVD和聚类用户最近邻的协同过滤混合推荐算法[J]. 计算机应用与软件, 2021, 38(05): 288-293.

[36]Geetha G, Safa M, FancyC, et al. A Hybrid Approach using Collaborative filtering and Content based Filtering for Recommender System[J]. Journal of Physics: Conference Series, 2018, 1000(1).

[37]Yadav P, Tyagi S. Hybrid fuzzy collaborative filtering: an integration of item-based and user-based clustering techniques[J]. International Journal of Computational Science and Engineering, 2017, 15(3-4): 295-310.

[38]刘峰, 王宝亮, 邹荣宇, 等. 基于随机游走的网络表示学习推荐算法[J]. 计算机工程, 2021, 47(09):90-96+105.

[39]Yang J H, Chen C M, Wang C J, et al. HOP-rec:High order Proximity for Implicit Recommendation[C]//Proceedings of ACM Conference on Recommender Systems.2018: 140-144.

[40]李盼, 解庆, 李琳, 等. 知识增强的图神经网络序列推荐算法研究[J/OL]. 计算机工程, 2022 (07): 1-12.

[41]He X, Deng K, Wang X, et al. Lightgcn:Simplifying and powering graph convolution

network for recommendation[C]//Proceedings of the 43rd international ACM SIGIR conference on research and development in information retrieval. 2020:639-648.

[42]吴成钢, 杨光, 张翔, 等. 推荐系统的应用及其安全性研究[J]. 信息网络安全, 2011 (08): 69-71.

[43]李宇琦, 陈维政, 闫宏飞等. 基于网络表示学习的个性化商品推荐[J].计算机学报, 2019, 42(8): 1767-1778．

[44]胡琪, 朱定局, 吴惠粦, 等. 智能推荐系统研究综述[J]. 计算机系统应用, 2022, 31 (04): 47-58.

[45]Kant S, Mahara T. Merging user and item based collaborative filtering to alleviate data sparsity[J]. International Journal of System Assurance Engineering & Management, 2018, 9(1): 1-7.

[46]Liu L, Lecue F, Mehandjiev N. Semantic content-based recommendation of software services using context[J]. ACM Transactions on the Web(TWEB), 2013, 7(3): 1-20.

[47]Lian Z, Hui C. Research on recommendation algorithms in web of things system[C]//47th International Conference on Intelligent Computation Technology and Automation, Changsha, 2014: 569-572.

[48]李忠俊, 周启海, 帅青红. 一种基于内容和协同过滤同构化整合的推荐系统模型[J]. 计算机科学, 2009, 36(12): 142-145．

[49]An Q. A Novel Recommendation Algorithm Considering Average Similarity and User-based Collaborative Filtering[J]. Mathematical Modelling of Engineering Problems, 2019, 6(3).

[50]魏甜甜, 陈莉, 范婷婷, 等. 结合项目流行度加权的协同过滤推荐算法[J].计算机应用研究, 2020, 37(03): 676-679.

[51]Li S. Research on User-item Rating based on Collaborative Filtering Algorithm[J]. International Journal of u- and e- Service, Science and Technology, 2016, 9(2).

[52]荣辉桂, 火生旭, 胡春华, 等 .基于用户相似度的协同过滤推荐算法[J]. 通信学报, 2014, 35(02): 16-24.

[53]于金明, 孟军, 吴秋峰. 基于改进相似性度量的项目协同过滤推荐算法[J]. 计算机应用. 2017, 37(5): 1387-1391+1406.

[54]Jena K K, Bhoi S K, Mallick C, et al. Neural model based collaborative filtering for movie recommendation system[J]. International Journal of Information Technology, 2022: 1-11.

[55]Meng X, Liu S, Zhang Y, et al. Research on social recommender systems [J]. Journal of Software, 2015, 26(6): 1356-1372.

[56]Rodrigues C M, Rathi S, Patil G. An Efficient System Using Item & User-Based CF Techniques to Improve Recommendation[C]. 2016 2nd International Conference on Next Generation Computing Technologies (NGCT), 2016: 569-574.

[57]Tang X, Zhou J . Dynamic Personalized Recommendation on Sparse Data[J]. IEEE Transactions on Knowledge & Data Engineering, 2013, 25(12): 2895-2899.

[58]Liu J, Dolan P, Pedersen E R. Personalized news recommendation based on click behavior[C]//International Conference on Intelligent User Interfaces,February 7-10, 2010, Hong Kong, China. DBLP, 2010: 31 -40.

[59]吴毅涛, 张兴明, 王兴茂, 等.基于用户模糊相似度的协同过滤算法[J]. 通信学报, 2016, 37(01): 198-206.

[60]Tahmasebi F, Meghdadi M, Ahmadian S, et al. A hybrid reco mmendation system based on profile expansion technique to alleviate cold start problem[J]. Multimedia Tools and Applications, 2020, 80: 2339-2354.

[61]Lumauag R G, Sison A M, Medina R P. An enhanced recommendation algorithm based on modified user-based collaborative filtering[C]//IEEE 4th International Conference on Computer and Communication Systems (ICCCS), Singapore, 2019: 198-202.

[62]Gazdar A, Hidri L.A new similarity measure for collaborative filtering based recommender systems[J]. Knowledge-Based Systems, 2020, 188: 105058.

[63]薛亮, 徐慧, 冯尊磊, 等.一种改进的协同过滤的商品推荐方法[J]. 计算机技术与发展, 2022, 32(07): 201-207.

[64]Lemire D, Maclachlan A. Slope one predictors for online rating-based collaborative filtering [C]//Proceedings of the 2005 SIAM International Conference on Data Mining.

Society for Industrial and Applied Mathematics, 2005: 471-475.

[65]陈梅梅, 董晨光, 王淇, 等. 兼顾用户话语权的改进加权Slope-One协同过滤推荐[J]. 小型微型计算机系统, 2022, 43(09): 1814-1819.