Prof. Ajay Jaiswal
Prestige Institute of Engineering Management and Research, Indore, Madhya Pradesh
452010, India
Abstract
Effective software-related cost estimation is paramount in decision-making. Estimating is the macro activity that is part of project methodology and allows for the effective delivery of projects. This is useful in project management because it assists with implementing the necessary tasks. Pretty much the discussed parameter helps in the optimization of resources in relation to the requirements for accomplishing project scope. There are several important factors that encompass software projects, including time, resources, human resources, infrastructure and materials, finance, and risk. In case the cost estimate is lower than required, the time for the development of the project will be longer and more expensive. The scope for waste of resources has been exaggerated. Artificial intelligence is a fusion of machine learning and deep learning to produce smart systems capable of posing solutions to problems. Software effort estimation assists in constructing the objectives, which include planning, scheduling, and budgeting for a project. Different prediction trials mentioned above, which were expert opinion-based, analogy-based estimates, regression estimations, categorization strategies, and deep learning algorithms, were suggested as predictors of type of endeavors. Among the evaluation metrics discussed were Mean Absolute Error, Root Mean Squared Error, Mean Square Error, and R-squared. Therefore, estimation has and will take a significant role in risk prevention measures in the future. Metrics for assessment will be used in many assessments. After this, other studies intend to explain the reasons why software developer cost modeling can be very beneficial in light of LSTM (Long Short-Term Model) and CNN (Convolutional Neural Network) prospects introduced throughout the research. This method allows for solving intricate tasks with multiple dependencies in an ever-changing environment by using ML (Machine Learning) and DL (Deep Learning) technologies. Further studies reveal that the most common deep learning architecture in these studies was convolutional, and relatively little application was deep learning.
Learn More