Enhancing Software Effort Estimation:A Hybrid Model Combining LSTM and Random Forest  

作　　者：Badana Mahesh Mandava Kranthi Kiran 

机构地区：[1]Department of Computer Science and Engineering,GITAM Deemed to be University,Vishakhapatnam 561203,India [2]Department of Computer Science and Engineering,Anil Neerukonda Institute of Technology&Sciences,Vishakhapatnam 531162,India

出　　处：《Journal of Harbin Institute of Technology(New Series)》2024年第4期42-51,共10页哈尔滨工业大学学报（英文版）

摘　　要：Effort estimation plays a crucial role in software development projects,aiding in resource allocation,project planning,and risk management.Traditional estimation techniques often struggle to provide accurate estimates due to the complex nature of software projects.In recent years,machine learning approaches have shown promise in improving the accuracy of effort estimation models.This study proposes a hybrid model that combines Long Short-Term Memory(LSTM)and Random Forest(RF)algorithms to enhance software effort estimation.The proposed hybrid model takes advantage of the strengths of both LSTM and RF algorithms.To evaluate the performance of the hybrid model,an extensive set of software development projects is used as the experimental dataset.The experimental results demonstrate that the proposed hybrid model outperforms traditional estimation techniques in terms of accuracy and reliability.The integration of LSTM and RF enables the model to efficiently capture temporal dependencies and non-linear interactions in the software development data.The hybrid model enhances estimation accuracy,enabling project managers and stakeholders to make more precise predictions of effort needed for upcoming software projects.

关 键 词：software effort estimation hybrid model ensemble learning LSTM temporal dependencies non⁃linear relationships 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]