Semnan University Press Journal of Modeling in Engineering 2008-4854 16 54 2018 09 23 A Novel Hybrid Model to Increase the Accuracy of Case Based Reasoning Method in Software Development Effort Estimation A Novel Hybrid Model to Increase the Accuracy of Case Based Reasoning Method in Software Development Effort Estimation 39 52 3382 10.22075/jme.2017.6140. FA Mozhdeh Sabbagh Nezhad Amid Amid Khatibi Bardsiri Journal Article 2017 01 01 Nowadays the effort estimation of software development is crucial in Software projects management. Not only have the accurate estimate of cost help customers and investors, but also it will be effective in rational decision-making in the implementation and management of software projects. Various estimation models have been invented and used so far. Many of the current effort estimation approaches are adopted by collecting data from previous projects. Case-based reasoning (CBR) is one of the successful techniques of effort estimation in software projects. This method alone is not very accurate, a defect which can be corrected by creating hybrid models. In this paper, CBR was combined with two separate metaheuristic algorithms including particle swarm optimization (PSO) and the firefly algorithm to propose a new hybrid model. Then the performance of the proposed model was evaluated. According to the results of the proposed model on Cocomo, Albrecht and Maxwell datasets, the firefly algorithm showed an acceptable performance. Nowadays the effort estimation of software development is crucial in Software projects management. Not only have the accurate estimate of cost help customers and investors, but also it will be effective in rational decision-making in the implementation and management of software projects. Various estimation models have been invented and used so far. Many of the current effort estimation approaches are adopted by collecting data from previous projects. Case-based reasoning (CBR) is one of the successful techniques of effort estimation in software projects. This method alone is not very accurate, a defect which can be corrected by creating hybrid models. In this paper, CBR was combined with two separate metaheuristic algorithms including particle swarm optimization (PSO) and the firefly algorithm to propose a new hybrid model. Then the performance of the proposed model was evaluated. According to the results of the proposed model on Cocomo, Albrecht and Maxwell datasets, the firefly algorithm showed an acceptable performance. https://modelling.semnan.ac.ir/article_3382_2de5095d2bc3e0423396ab65b7174d37.pdf