【摘要】 目的将基于遗传程序设计(GP)的数据建模方法应用于税收能力智能估测系统建模中。方法利用遗传程序设计来实现复杂函数的自动建模。结果改变了目前税收能力估测中采用的使用拟合粗糙、估测结果准确度差的线性模型进行曲线拟合和趋势预测的估测方法,通过优化初始种群的方法提高了GP算法的收敛率,实现了复杂函数的自动建模,并将该方法应用到税收能力智能估测原型系统(TCES)的建模中。结论采用GP得到的数学模型比传统方法使用的数学模型拟合精度更高,估测误差更小,更适于反映地区税收能力的实际情况。更多还原

【Abstract】 Aim To apply Genetic Programming(GP) to the Modeling of TCES(Tax Capacity Estimation System).Methods Applying Genetic Programming to the automatic modeling of complex functions.Results This method can obviously improve the traditional method of estimation and economic curve fitting which could only obtain rough fitting curves and unsatisfactory estimation results in the Estimation of Tax Capacity. It presents an improvement over the initial population in order to strengthen the convergence performance and realizes the automatic modeling of complex function,as well as its corresponding test processing and modeling process for TCES by means of the method.Conclusion The numerical results indicate that the mathematical models set up by GP is more accurate than the traditional models.It′s more suitable to reflect the fact of regional tax capacity.更多还原