In order to obtain the data corresponding to certain condition or status from mass test data quickly and accurately, the problems existing in conventional time-sequence search methods were analyzed, and a fast search approach for multi-channel mass data was proposed based on optimal entropic matching. In this approach, the conventional " time matching" is replaced by " status matching". The approach depends mainly on entropy testing to need few parameters, have less reliance on neighborhood time and largely reduce the error caused by time synchronization between different channels. Furthermore, it has a big search scale and a fast search speed. The result shows that in certain condition, the search speed can be increased by thousands of times with an improved matching precision, and the relative error of every main parameter is less than 0.05.%为快速、准确地从海量试验数据中获取所需的与某一工况或状态对应的数据,通过分析常规时序搜索方法存在的问题,根据信源熵的原理和特点,提出了一种基于最佳熵匹配的多通道海量数据快速搜索方法.该方法用“状态匹配”代替常规的“时序匹配”,以检测熵值为主,所需参数少,对邻近时间的依赖性小,降低了因通道间时间同步产生的误差,搜索步伐大,速度快.结果表明,对于一些特定的工况,搜索速度可提高数千倍,匹配准确度提高,主要参数的相对误差均在0.05以下.
展开▼
