Causal Model of Variables by Information Theory
Journal Title: International Journal of the Computer, the Internet and Management - Year 2018, Vol 26, Issue 1
Abstract
The objective of this research was to find out the relationship or dependency of independence random variables. Data that was used to experiment in this technique was selected from CM-1 Dataset provided by Promise. There were many independence variables that were composed to a binary logistic regression model to predict if software project should be either defect or free of bugs. All of these independence variables (xi) were used to predict dependence variable (y – defect or not defect). Chi- square test was used to choose which independence variables were related to dependence variable. Many independence variables were not significance related or dependence to class variable Y. After that, these unrelated independence variables were tried to figure out significance their relation and either cause or effect direction with all related independent variable (with Y). Information theory was used to find out direction of either cause or effect between two independence variables (one variable that related to Y and other that was not related to Y). These relations were considered their relation importance by correlation significance test-t-Test at significance level ∝ 0.05.
Authors and Affiliations
Aryut Ruangtong, Nalinpat Porrawatpreyakorn, Somchai Prakancharoen
Keywords
Related Articles
- EP ID EP598685
- DOI -
- Views 174
- Downloads 0
Effects of Induction Motor Using Unbalance Voltage
This article presents the effects of inductive motor was operated under imbalanced voltage condition. It was functional test of electric motor at field work by analyzing effect to motors in 3 aspects as followed: tempera...
Synthesis of Indicators of Scientific Literacy Components for Junior High School Students: Via Online System
The objective of this research study aimed to synthesize indicators of scientific literacy components for junior high school students: via online system. The source of data is related documents and research studies assoc...
Study of Brand Love as Consequence of Virtual Brand Community Engagement
Brands today face a new challenge as consumers take the communications into their hands using social media to connect more among themselves. The traditional view that consumers are the passive recipients of communication...
A Guideline for Forest Development to Promote Mini-Hydropower: a Case Study in 4 Community Forests in Lampang Province
Healthy community forests can benefit the society by releasing water mass for small scale hydropower plants. This study has an objective to obtain a guideline for accelerating development of community forests that ideall...
Automatic Course Planning System Using Rule-Based Ontological Knowledge Base
This research presents how ontology combined with semantic rules can be used to develop an automatic planning system. The question is how to match a course plan for students in a class, because there are various types of...