[目的/意义] Web 2.0时代，在线评论质量参差不齐和过载问题十分严重，人们从中获取有价值内容的认知成本越来越高。探究以信息推荐方式解决评论过载的有效方案，以提升网络信息利用率和信息服务质量。论文提出的评论排名推荐方案关注评论的信息质量，更强调对用户的个人信息需求的满足。[方法/过程] 研究运用概率主题模型，引入词向量构建主题空间下的用户模型和评论模型，通过将其纳入评论感知效用评测系统，实现融合用户兴趣和评论质量的评论推荐，推荐效果通过系统实验予以检测。[结果/结论] 实验结果表明，评论信息质量和用户个体的信息需求，共同作用于用户对评论感知效用的满意度；推荐策略实现了二者的有机融合，三组不同推荐模式下的评测效果显示，相较于单纯的"兴趣推荐"和"效用推荐"，"融合推荐"综合满意度得分最高。
[Purpose/significance] In the Web2.0 era, the quality of online reviews is uneven and overloaded; the cognitive cost of getting valuable content from them is increasing. In the paper, we proposed a review recommendation ranking scheme which focuses on the information quality of reviews and emphasizes more on the satisfaction of users' personal information needs. [Method/process] The probabilistic topic model was adopted in this study, user-profile and review model based on the topic model were generated by employing the Word2Vector. By incorporating them into review helpfulness evaluation system, the comment recommendation which integrates user interest and comment quality was realized. The models and related methods were tested and evaluated by a serial of systematic experiments. [Result/conclusion] The results indicate that both information quality and individual information interests are of influence on review perceived helpfulness. The "combining-specific" recommendation strategy, integrating the two factors effectively, performs better than the "interest-specific" and the "utility-specific" recommendation method. From the perspective of information service, the "combining" strategy should be with the high-priority for review recommendation.
[1] 郭顺利,张向先,李中梅.面向用户信息需求的移动O2O在线评论有用性排序模型研究——以美团为例[J]. 图书情报工作, 2015, 59(23):85-93.
[2] 张艳丰,李贺,彭丽徽, 侯力铁. 基于情感语义特征抽取的在线评论有用性分类算法与应用[J]. 数据分析与知识发现, 2017, 1(12):74-83.
[3] 王忠群,吴东胜,蒋胜, 等. 一种基于主流特征观点对的评论可信性排序研究[J]. 数据分析与知识发现, 2017, 1(10):32-42.
[4] 吴璠,王中卿,周夏冰,等. 基于文本和用户信息的在线评论质量检测[J]. 中文信息学报, 2019, 33(9):107-114, 140.
[5] 祝琳琳,李贺,刘金承,等. 在线评论信息质量感知评价指标体系构建研究[EB/OL].[2020-12-12]. https://kns.cnki.net/KXReader/Detail?TIMESTAMP=637449586862656250&DBCODE=CJFQ&TABLEName=CAPJLAST&FileName=QBLL20201112002&RESULT=1&SIGN=MYZ8niwHlIVHm8g%2f8tmHWEKEjGs%3d.
[6] 李健,张军,苑清敏,等. 在线商品评论对消费者效用的改进分析——基于信息质量和消费者满意度理论视角[J]. 情报科学, 2018, 36(7):137-144.
[7] BEN-ABDALLAH E, BOUKADI K, HAMMAMI M. Personalized cloud service review ranking approach based on probabilistic ontology[C]//Proceedings of 22nd international conference on business information systems. Seville:Springer, 2019:50-61.
[8] 姜霖, 张麒麟. 基于评论情感分析的个性化推荐策略研究——以豆瓣影评为例[J]. 情报理论与实践, 2017, 40(8):99-104.
[9] CHEN L, CHEN G, WANG F. Recommender systems based on user reviews:the state of the art[J]. User modeling and user-adapted interaction, 2015, 25(2):99-154.
[10] ESPARZA S G, P.O'MAHONY M, SMYTH B. Effective product recommendation using the real-time web[C]//Proceedings of the 30th SGAI international conference on innovative techniques and applications of artificial intelligence. Cambridge:Springer, 2010:5-18.
[11] 耿立校,晋高杰,李亚函,等.基于改进内容过滤算法的高校图书馆文献资源个性化推荐研究[J]. 图书情报工作, 2018,62(21):112-117.
[12] ZHANG W, DING G, CHEN L, et al. Generating virtual ratings from Chinese reviews to augment online recommendations[J]. ACM transactions on intelligent systems and technology, 2013, 4(1):1-17.
[13] MUSAT C, LIANG Y, FALTINGS B. Recommendation using textual opinions[EB/OL].[2020-12-26]. https://www.researchgate.net/profile/Claudiu-Musat/publication/262348054_Recommendation_using_textual_opinions/links/57480c6308ae2301b0b8771f/Recommendation-using-textual-opinions.pdf.
[14] 张宜浩,朱小飞,徐传运,等. 基于用户评论的深度情感分析和多视图协同融合的混合推荐方法[J]. 计算机学报, 2019, 42(6):1316-1333.
[15] 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.
[16] 张炎亮,张超,李静. 基于动态用户画像标签的KNN分类推荐算法研究[J]. 情报科学, 2020, 38(8):11-15.
[17] BAO Y, FANG H, ZHANG J. TopicMF:Simultaneously exploiting ratings and reviews for recommendation[C]//Proceedings of the 28th AAAI conference on artificial intelligence. Quebec:AI Access Foundation, 2014:2-8.
[18] FENG S, CAO J, WANG J. et al. Recommendations based on comprehensively exploiting the latent factors hidden in items' ratings and content[EB/OL]:[2020-12-26]. https://dl.acm.org/doi/pdf/10.1145/3003728.
[19] ZHENG L, NOROOZI V, YU P S. Joint deep modeling of users and items using reviews for recommendation[C]//Proceeding s of the 10th ACM international conference on web search and data mining. Cambridge, UK, 2017:425-434.
[20] SEO S, HUANG J, YANG H, et al. Representation learning of users and items for review rating prediction using attention-based convolutional neural network[EB/OL].[2020-12-26]. https://doogkong.github.io/2017/papers/paper8.pdf.
[21] CHEN C, ZHANG M, LIU Y, MA S. Neural attentional rating regression with review-level explanations[C]//Proceedings of the 2018 WWW conference on World Wide Web. Lyon, 2018:1583-1592.
[22] 冯兴杰,曾云泽,崔桂颖. 基于近邻用户评论的推荐辅助网络[J]. 计算机应用研究, 2020, 37(10):2956-2960.
[23] BLEI D M, NGA Y, JORDAN M I. Latent dirichlet allocation[J]. Journal of machine learning research. 2003, 3(4/5):993-1022.
[24] 中文预训练模型[EB/OL].[2020-12-20]. https://github.com/Embedding/Chinese-Word-Vectors.
