Abstract

In the software industry, software effort estimation is an essential activity since it has the potential to dramatically improve the quality of the parameters that the team is working with. The failure of a company to pay attention to crucial aspects during the process of developing and delivering software frequently results in major financial losses for the company. Some of the most conventional methods for design and analysis, such as the COCOMO model and others, provide an estimate of the costs and the amount of effort involved in an abstract manner. All of the requirements for the project management process will not be satisfied by this method. In response to this, a number of authors have proposed a methodology for estimating the amount of effort required to develop software that makes use of artificial intelligence and machine learning. A large number of models are based on a single neural network model approach, which is not always sufficient to produce satisfactory outcomes. Consequently, the assembly of many deep learning models in order to improve the estimation of the amount of work required by software is a suitable choice. In addition, a few of the authors go into further information about this topic by presenting a multitude of innovative concepts that, in the end, open up the possibilities of the software effort estimating process. As a result, the system is proposed to estimate the software efforts by employing the Support vector machine to classify the data. On the basis of this classification protocol, LSTM and ANN Neural network are implemented. The values of the predictions that were obtained are made available to the hybrid model so that it can make a decision on the estimation of the software efforts.