Using an Artificial Neural Network for Predicting Embedded Software Development Effort
Kazunori Iwata, Yoshiyuki Anan, Toyoshiro Nakashima, Naohiro Ishii · 2009
In this paper, we establish an effort prediction model using an artificial neural network (ANN) for complementing missing values. We add missing values to the data via collaborative filtering using the method of Tsunoda et al.'s method. In addition, we perform an evaluation experiment to compare the accuracy of the ANN model with that of the MRA model using Welch's t-test. The results show that the ANN model is more accurate than the MRA model, since the mean errors of the ANN are statistically significantly lower.