Article Sidebar
Download
Main Article Content
Abstract
Ridge polynomial neural network (RPNN) originally proposed by Shin and Ghosh, constructed from an increase in the number of order pi-sigma neurons (PSN). RPNN maintains faster learning, a stonger mapping of single layer of higher order neural network (HONN) and avoids heavy loads due to the increasing number of inputs. Chaos optimization algorithm is used by utilizing the logistic equation which is sensitive to initial conditions, so that the movement of chaos can be changed in any circumstances in partcular scale within regularity, ergodic and maintain the diversity of solutions.Software cost of prediction is a very important process in software development. The ignorance of cost prediction of the software will cause in poor quality of software produced as well as creates a new problem in the software financial system. To anticipate the occurrence of errors in software cost of estimation, it is necessary to develop a calculating method or prediction cost of software so that software cost which predicted be more optimal. This research will develop a way by using the Chaos Optimization Algorithm in Ridge Polynomial Neural Network to predict the cost of software.Networking process training is using RPNN while of the initial value loads the searching and networking biases are using chaos optimization algorithm. The structure used consists of six (6) layer neurons and one (1) neuron layer output. This research uses data from NASA project undertaken between 1980 untill 1990. The results of this research demonstrates that the proposed algorithm can be used for prediction