基于时序相关性的信息效用评价模型研究  

Research on Evaluation Model of Information Utility Based on Timing Correlation

作　　者：杜晓艳[1] 

机构地区：[1]深圳大学图书馆,广东深圳518060

出　　处：《郑州航空工业管理学院学报》2016年第4期128-131,共4页Journal of Zhengzhou University of Aeronautics

摘　　要：事务处理中的时间相关性数据蕴涵着对预测与决策具有情报效用的信息,可以通过构建一定的特征模式对其效用大小进行评价。文章对时变序列数据的相关性特征进行了分析,研究了基于时序与波动变化的特征抽取模式;在时变序列的差分及相对变化基础上,进一步综合利用偏态、离散、相关等特征构建了影响变量与响应变量之间的相关度计算模型。该模型可降低效用评价计算复杂度,增强评价的可靠性并可用于模型预测。The time- related data obtained by processing transaction contain the infoimation with different utility values for supporting prediction and decision -making, and can be used to evaluate utility magnitudes by establis- hing definite feature patterns. In this paper, the correlations of are analyzed, and the extraction patterns of feature of time series data are studied based on timing and fluctuation characteristics. Built on the difference and relative change of time - varying series, the computation module of correlation degree is constructed between influence vari- ables and response variables by comprehensive utilization of partial state, discrete, correlation, etc.. The algorithm reduces the computational complexity ,enhances the reliability of the evaluation, and can be applied to the predic- tion.

关 键 词：时序特征 相关性 特征模式 信息效用 信息资源 

分 类 号：G271.4[文化科学—档案学]