Vague集在软件成本估算中的应用

类比估算方法是软件成本估算的常用方法之一,而估算中选择的相似度函数则是估算结果准确有效的关键.由于Vague集可以较好地处理数据中的未知信息与不确定性,因此可使用Vague集来提高软件成本估算的准确性,为项目管理人员提供更加科学的预算保障.为此充分考虑未知度的因素,提出了一个新的基于未知度的Vague集相似度量方法,结合Vague集模糊熵与熵值法来确定不同指标的权重,以获取不同项目之间的相似度.最后将该方法应用在Desharnais数据集进行成本估算,实验结果表明,该相似度量方式较现有方法能够进一步降低估计过程中产生的误差.

Application of Vague sets to software cost estimation

The analogy-based effort estimation is one of the commonly used methods for software cost estimation,and the selection of the similarity function is the key to the accuracy of estimation. Because the Vague sets can better deal with the unknown information and uncertainty in data processing,they can be used to improve the accuracy of the final result of the software cost estimation,providing a more scientific,standardized and systematic budget for the project management personnel. Toward this end,considering the factor of the unknown degree,this paper has proposed a new similarity measure based on the unknown degree between Vague sets,combining the fuzzy entropy of Vague sets and the entropy value method to determine the weights of different indexes in order to obtain the similarities between the properties of the projects. Finally this method is applied to the Desharnais dataset for cost estimation. The experimental results show that the proposed method can further reduce the estimation error compared with the existing method.