论文部分内容阅读
数据深度利用是税收信息化发展的必然趋势。税收数据本身是一种相对复杂的数据,具有多因子、高噪声、随机性、非线性、非高斯分布和不均匀分布等特性,常规的数理统计方法往往不能适应日益增长的数据利用需求。税收数据建模将数理统计、模式识别、人工修正、机器学习等方法融合在一起,对税收数据进行深度挖掘,将隐蔽的、有价值的规律挖掘出来。税收数据模型主要有分行业建模和分税种建模两大类,相比较而言,本文认为对于基层税务机关来说,结合本地区行业特点搭建行业税收数据模型,更具有针对性和可操作性。本文结合本单位搭建水泥行业税收数据模型实践,探索建立行业税收数据模型的一般方法。
The use of data depth is the inevitable trend of the development of tax information. Tax data itself is a relatively complex data with many factors, such as multiple factors, high noise, randomness, non-linearity, non-Gaussian distribution and uneven distribution. The conventional mathematical statistics methods are often unable to meet the increasing demand for data utilization. Tax data modeling Combining mathematical statistics, pattern recognition, manual correction, machine learning and other methods, the tax data is deeply tapped, and the hidden and valuable laws are excavated. Tax data model is divided into two major categories of industry modeling and tax-sharing modeling, comparatively speaking, this paper believes that for the tax authorities at the grassroots level, combining the characteristics of the industry in the region to build industry tax data model is more targeted and operational Sex. In this paper, the establishment of the cement industry tax data model practice to explore the establishment of industry tax data model of the general method.