Analisis Metode Bayesian Network and Multivariet Adaptive Regression Splines (MARS) Untuk Memprediksi Object-Oriented Software Maintainability

Rosyid, Harunur (2013) Analisis Metode Bayesian Network and Multivariet Adaptive Regression Splines (MARS) Untuk Memprediksi Object-Oriented Software Maintainability. UMMPress, Gresik.

[img] Text
Anlisis Metode Bayesian.pdf
Download (3MB)

Abstract

Object-oriented metrics that have been proposed needs to be validated, One possible way to validate metrics is to do a statistical analysis of the size metrics and system maintainability. The problem is the number of metrics that are used to predict the maintainability of 00 (Object Oriented) as long as it was not known metric that has the best prediction accuracy, so the analysis is needed tofind a better metric. Evaluation model, is used to introduce the criteria for evaluating the accuracy of model predictions. Bayesian network is used as maintainability prediction model, one that is connected to the node CHANGE ten metric variables, Bayesian networks models that predict its output estimates in this paper, uses the interval estimation point. This is because the interval estimate is difficult to compare with other models suitable for the sample test used to test and compare the MARS models and other models. This paper evaluates and compares the 00 software maintainability prediction models quantitatively, by using the prediction accuracy with a step size measures: absolute residuals (Ab.Res.), magnitude of relative error (MRE) and Pred measures. Analysis results can be concluded that the matrix model of MARS can effectively predict the maintainability of 00 software systems. This is better than the matrix model of Bayesian network as the absolute value of residuals is significantly higher.

Item Type: Other
Uncontrolled Keywords: metric, Object Oriented, maintainability, Bayesian network, MARS. absolut residual. admistration
Subjects: Engineering > Electronical Engineering
Engineering
Divisions: Faculty of Engineering > Informatics Engineering Study Program
Depositing User: Admin Admin Admin
Date Deposited: 20 May 2025 09:32
Last Modified: 20 May 2025 09:32
URI: http://eprints.umg.ac.id/id/eprint/13749

Actions (login required)

View Item View Item
UMG REPOSITORY is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.