Abstract: Existing evaluation methods do not consider the time decay of behavior history records, and there are some small number of malicious users with less frequency, low-value access for higher confidence. In order to resolves trust issues decay over time, this paper proposes a grid entity behavior evaluation method based on sliding window by constructing a time correction function. This method can make the calculation of trust degree more accurate, timely and objectively reflect the real behavior of entity, and it provides the correct basis for the system in making decision about follow-up behaviors of the entity.

Key words: grid entity behavior, trust evaluation, time correction function, sliding window

摘要： 现有信任评估方法没有考虑行为历史信息记录的时间自衰减性，并存在一些恶意用户以少次数低价值访问换取高信任的情况。为此，在构造时间校正函数基础上，提出一种基于滑动窗口的网格实体行为信任评估方法，能准确计算出实体信任度，客观反映实体行为的真实性，为系统对实体的后续行为控制提供决策依据。

关键词: 网格实体行为, 信任评估, 时间校正函数, 滑动窗口

CLC Number: 

• TP391