软件成本组合估算模型的3种权重系数确定法

摘要

综合互补神经网络和向量机的优缺点，在研究并构建基于 RBF神经网络和 RVM的软件成本组合估算模型的基础上，重点应用熵值法、二次规划法和方差-协方差法以确定该组合估算模型的权系数，最后采用 COCOMO数据库数据为输入，通过实验验证3种权系数确定方法的合理性和精确性。%This article complementarily synthesizes the advantages and disadvantages of neural networks and SVM,and on the basis of researching and building software cost estimation model based on RBF neural network and RVM,and by mainly using Hawkeye entropy method,quadratic programming,and the variance-covariance method to determine the weight coefficient of the combined estimation model,and finally,using COCOMO database data as input,verifies by experiments the reasonableness and accuracy of the three methods for determining the weight coefficients.