个性化推荐中的“过滤气泡”现象相关研究综述

A Summary of Studies on the “Filter Bubble” Phenomenon in Personalized Recommendations

个性化推荐的广泛开展在为用户采纳信息提供便利的同时,也产生"过滤气泡"效应带来信息窄化、内容质量下降等危害。从"过滤气泡"影响过程、产生原因、应对策略、业内尝试等对相关研究进行综述,指出"过滤气泡"有正反两方面影响,是推荐算法和人为因素复合作用的结果,可以从提供多样性尝试、增加信息偶然性、承担算法责任等方面进行策略研究或实践尝试,并对研究趋势、特别是图情领域的研究优势和注重方向进行展望。

The extensive development of personalized recommendations facilitates the adoption of information for users, as well as the "filtering bubble" effect, which leads to the narrowing of information and the degradation of content quality. The article reviews the related research from the aspects of "filtering bubbles", the causes, coping strategies, and industry attempts. It points out that "filtering bubbles" have both positive and negative effects, which is the result of the combination of recommended algorithms and human factors. We can study the strategy or practice from the aspects of providing diversity, increasing information contingency and assuming algorithmic responsibility, and look forward to the research trend, especially the research advantage and focus on the direction in the field of library and information science.