Novel approach for software reliability based on immune technique filters

Abstract

In this thesis an innovative model Software Reliability Based on Immune Technique SRBIT and Neural Network Classification Methods NNCM is proposed to ensure the software reliability with High Efficiency Immune Filters HEIF and Data Normalization of software metrics With the newlineavailability of cost effective technology about 80 of software systems are running on legacy platforms Reliability assessment of upgraded legacy systems is an important problem in any company s IT infrastructure Some parts of the code used in the original design of such systems might be currently being discontinued Trustworthy reliability assurance after an upgrade with a new software component is needed which combines the evidence about the reliability of the new component with the knowledge about the old system accumulated to date In these circumstances machine learning approach to reliability assessment is invaluable It is proposed to study data preprocessing on tupules to improve the classification of legacy modules using various data newlinemining techniques for practical reliability assessment Irregularities are widespread if the software has a large number of modules and often lead to erroneous conclusions with respect to data mining and statistical analysis newline