模糊粗糙集理论在纳税评估中的应用

An Application of Fuzzy Rough Set in Tax Assessment

纳税评估是税收管理工作的一个重要组成部分,纳税评估指标体系的建立是纳税评估工作的一个重要环节。关键是评估效率和准确性。为了提高纳税评估的业务效率,准确地对纳税人的申报税款情况进行评价,利用模糊粗糙集理论,选用不同的离散区间化方法,对选取的评估指标进行属性约简,最后采用支持向量机机器学习的方法对约简属性进行评价,经过实验比较分析后,得出一种新的纳税评估指标体系,通过实例验证。实验结果表明指标体系能很好地提高纳税评估预测性能。

Tax assessment plays an important role in taxation management in China.An important part of tax assessment is to establish an index system for tax assessment with fuzzy rough set to reduce the complexity of the selected index feature.The SVM is applied to classify the taxpayer,and a new index system for tax assessment can be established. The experimental results show that index system performs well in both data classification accuracy and predictive accuracy.