Modeling and Representation of Software Framework Cost Estimation Using Convolution Neural Networks

Software cost estimation is the important activity in the development of projects. It aids project managers and software developers in resource planning and management. The cost of constructing a software design was predicted using the convolution neural network (CNN) technique and compared to the conceptual design stage. The goal of cost estimates is to properly analyze necessary assets and timeframes for software enhancement projects, and it covers a wide range of topics, including calculation of the dimensions of software designs to be created, assessment of the work necessary, and assessment of the project's cost. The accuracy of development cost projection has an influence on the total project life cycle. To adapt the framework with its parameters to estimate Software Development (SD) effort, a multilayer feed forward neural network (FFNN) is proposed. The back propagation (BP) learning approach is utilized for training the network by repeatedly processing training samples by comparing the model prediction with the original effort. The COCOMO framework employs a multi layer FFNN. The COCOMO dataset is updated to test and develop the system. The purpose of this work is to provide a quantitative measure that can be used in both the present and suggested models.