Functional Equivalence between Radial Basis Function Neural Networks and Fuzzy Analogy in Software Cost Estimation

We show in this paper the equivalence between the radial basis function networks and fuzzy analogy in the field of software cost estimation. We prove that under weak conditions, the three layers of RBFN are functionally equivalent to the three steps of fuzzy analogy. This functional equivalence implies that advances in each literature, such new learning rules or new similarity measures, can be applied to both models directly. Furthermore, this equivalence can help us to provide a natural interpretation of cost estimation models based on RBFN.