An investigation of using Neuro-Fuzzy with software size estimation

Abstract

Neuro-Fuzzy refers to a hybrid intelligent system using both neural network and fuzzy logic. In this study, neuro-fuzzy is applied to a backfiring and categorical data size estimation model. Evaluation was conducted to determine whether a neuro-fuzzy approach improves software size estimations. It was found that a neuro-fuzzy approach provides minimal improvement over the traditional backfiring sizing technique.