The Research On Reliability Prediction Based On Orthogonal Layer Clustering Algorithm

The software reliability project may apply in any based on software product any edition, may start in any edition cycle. It is regarding a software product quality guarantee science appling in the software engineering entire life cycle. But the software reliability growth model plays an important role in this process. Through reasonable forecast we may obtain some positive software reliable target, which can control in the software process to carry on the correlation forecast well. It is well known, the software reliable forecast analysis is based on the massive flaw data, but through the static state trend analysis of massive software reliability growth model, the individual unusual data pointing to in the software reliability forecast process perturbation is extremely obvious and universal, and these data cannot the complete deference in the distribution curve. Has these phenomenon reason to include: Does not carry on the test in the analogy energy region selection test with the example to be able to obtain the different expiration data distribution; The software reliability testing input territory data distributions and the selection also affect in time domain expiration data distributed tendency and so on. In the software reliable process, the primitive supposition threshold excessively is high, but have representative also the spread ability big flaw data regarding cannot the uniform distribution cannot achieve the enormous collection, moreover regarding does not conform to this supposition request perturbation data to appear helplessly.This article discussed the limitations of the traditional model in software reliability analysis, presents a new prediction analysis method of software reliability model based on OLC algorithm, and the orthogonal cluster structure. The discussion of architecture of analysis is composed of three sections: first, the limitations of prediction analysis of traditional software reliability model are described. Secondly, the orthogonal cluster structure algorithm is introduced and the improved model based on this algorithm is presented. At last, the prediction analysis is implemented and data is analyzed. In the first section, based on the traditional J-M model, the process of prediction analysis of software reliability model and its relativeconception is introduced, the limitations of this method is discussed. In the second section, a method based on orthogonal cluster structure algorithm is presented in order to solve the problems raised in the traditional method and relative conceptions are also introduced. In the third section, the new algorithm is applied to the J-M model, and the traditional accumulated defective data is used in actual software reliability prediction analysis.This article which proposes the algorithm carrying on the software reliability in the improvement J-M model to forecast the analysis has solved the problem which the first part proposed, enhanced the forecast efficiency and the accuracy, the usability is strong. Simultaneously has very many places to be worth further improving and the thorough research, and finally discussed this architecture insufficiency and the forecast in this article.