【摘要】 【目的/意义】个性化信息推荐可能诱使用户受困“信息茧房”“过滤气泡”和“回音室”，诱发网络信息的“集中偏见”“流行度偏见”和“长尾效应”的损失等过度特化现象。梳理和探讨过度特化问题的表现形式、内在机制和应对策略对网络用户信息行为、个性化信息推荐研究，以及网络信息生态治理都具有重要的理论价值和现实意义。【方法/过程】梳理过度特化问题的相关概念与表现形式，分析过度特化问题的产生原因和内在机制，阐释过度特化问题的应对方法与策略，并提出未来研究方向。【结果/结论】过度特化问题根源于个体应对信息过载的选择性暴露行为，扩大化于信息生命周期，并在条件性行为反馈的服务机制、非透明性的推荐算法机制的助推下日益突出。过度特化问题的解决路径包含用户信息行为干预和推荐系统优化两类主要方法。【创新/局限】本文丰富和完善了过度特化的概念内涵，从信息系统、个体用户和信息三个方面系统梳理了过度特化相关表现形式、内在机制和实践路径，有利于推动过度特化问题的跨学科研究。更多还原

【Abstract】 【Purpose/significance】The personalized recommendation may cause users to be trapped in "information cocoon", "filtering bubble" and "echo chamber",thus contributing to the formation of over-specialization problems such as popularity bias,concentration bias,and the loss of the tail long-tail effect.It is of great theoretical and practical significance to sort out and summarize the manifestation,internal mechanism,and coping strategies of the over-specialization problems for the benign development of network public,network information ecology,and personalized information service.【Method/process】This paper introduces the related concepts and forms of over-specialization,hackles the causes,and analyzes the mechanism of over-specialization.Then it explains the coping strategies of the over-specialization from the perspective of online user behavior intervention and the reconstruction of the recommender system.Finally,future research directions of overspecialization researches are proposed.【Result/conclusion】The over-specialization originated from the individual’s selective exposure behavior and related cognitive psychological processes in response to information overload,and is magnified by the information life cycle under the effect of individual cognitive choices,then it has become increasingly prominent with the help of the service mechanism of conditional behavior feedback and the non-transparent recommendation algorithm mechanism.The solution to the problem of over-specialization includes "nudging" and "boosting" network user behavior intervention strategies and tools,as well as the construction of user-centered value orientation,a recommendation mechanism that integrates transparency,and over-specialization evaluation indicators.A recommended system process optimization method based on four aspects of the recommended optimization algorithm weighed against multiple indicators.【Innovation/limitation】This paper enriches and develops the concept of overspecialization,and systematically analyzes the phenomena,mechanisms,and practices related to overspecialization from three aspects:information systems,users,and information,to facilitate the interdisciplinary research on overspecialization.更多还原