Back Propagation Algorithm Based Model for Software Cost Estimation

Sheena Goyal, Sheilly Padda
2017 International Journal of Advanced Research in Computer Science and Software Engineering  
Software Cost Estimation is a tough task as it includes various variations in every step of implementation which creates problems in measuring the KLOC (Line of Source Code in thousands) for cost estimation. Software industry becomes very vast from decades but software cost estimation is still a big problem for which software industry is still suffering. Our proposed model is for tuning parameters of COCOMO model software cost estimation using Particle Swarm Optimization (PSO). We will be using
more » ... clustering methods to divide the data items into number of clusters. Once the data has been divided, it will be easier to implement Particle swarm optimization on each cluster. PSO is used for tuning the parameters of each cluster. Tuning stands for by continuously processing (here processing is applying PSO) in the data until the best value is found. The clusters and the tuned parameters will be trained on Neural Network by back propagation algorithm. Back propagation is done by calculating the equation once again using the better values found in our approach.
doi:10.23956/ijarcsse/v7i6/0278 fatcat:bbvdbuyxg5cy5fzehh3p3oliwe