Research of Matching Algorithm for Fault Data of Flight Query System

Software reliability data is the basis for software reliability evaluation, the validity of the data plays a crucial role in the accuracy of software reliability assessment. In general, the original data is directly used for reliability model, however it ignores repeatability, false alarm, incompleteness and other defects of the original data. These defects seriously affect the accuracy of reliability evaluation. In order to solve above problems, this paper presents a matching algorithm bases on sliding time window for fault data. Firstly, structured processing monitoring data, and generating alarm time series according to the alarm time sequence. Determining the width and sliding step of time window. Then cleaning the data of each time window, retaining the alarm key information, and generating alarm data. Further compressing the alarm data within time window so as to form fault data. Finally, switching fault data into reliability data. Experiments with flight query system as an example, generating two types of reliability data, and applying the data to the fault count model and the fault time interval model for reliability evaluation. The results indicate that the efficiency and feasibility of the matching algorithm bases on sliding time window for fault data is enhanced.