Enrichment of Fault Features by Forming Ml Hypothesis

Abstract

Modeling and optimization of applications of engineering sciences and technology advances in fault detection and diagnostics.  By utilizing past data one is to promote environmentally safe modeling approaches.  The software developers and users have found difficult to learn the software fault because softwares are developed using most of the learning algorithms. So, the developers needed some learning technique in order to prevent and identify the fault in pre-development.This will leads to the introduction of green engineering in software development. This paper examines and forms the hypothesis space for fault features classification in post release so as to form the learning technique to identify it in the development stage itself to reduce rework. This paper also checks the classification of input features that are to be relevant to the outcome to be predictedare not by using different hypothesis testing. Our result signifies the hypothesis space using machine learning for finding feature set of fault prediction feature set. Eight NASA PROMISE Repositories are used in this paper for the hypothesis testing.  This paper used to identify the best Hypothesis Testing for solving the feature selection problem in machine learning Hypothesis Space. Several performance measures are calculated and results of the experiment revealed that choosing chi-square hypothesis testing produces more relevant result for fault prediction feature set formation.