Tax payer industry classification-oriented label noise learning method

Abstract

The invention discloses a taxpayer industry classification-oriented label noise learning method, which comprises the following steps of: firstly, extracting text information and non-text information in taxpayer industry information, and respectively carrying out text embedding and non-text coding processing based on an XLNet text pre-training network and a coding technology to obtain feature information; secondly, constructing a TextCNN network used for taxpayer industry classification, determining the number of layers of the network, the shape of a convolution kernel and the input and output dimension of each layer according to the feature information and the target classification number, connecting the XLNet text pre-training network and the TextCNN network in series, and constructing an end-to-end training device by combining the noisy taxpayer industry label data as supervision; thirdly, estimating a condition transfer matrix based on an improved mixing proportion estimation method; and finally, learning network parameters in the training device, taking the condition transfer matrix as a linear layer behind the TextCNN network, realizing conversion from noise label prediction to real taxpayer industry label prediction, and carrying out taxpayer industry classification.

Description

technical field [0001] The invention belongs to the technical field of text classification with label noise, in particular to a label noise learning method for taxpayer industry classification. Background technique [0002] In recent years, the market economy has continued to prosper, the number of enterprises has increased, and the division of labor has been continuously refined. Accompanying this, the upgrade and further construction of the tax system has become an urgent need. [0003] Taxpayer industry classification is a prerequisite for determining taxpayer policies and preferential policies, and an important link in tax collection. At present, my country divides the taxpayer industry into 20 categories and 97 categories. Due to the large number of categories, the traditional manual classification method requires a lot of human resources, and is limited by the professional knowledge and experience of the classifier, which inevitably introduces classification errors, ...

AI Technical Summary

Problems solved by technology

However, the traditional mixed ratio estimation method is only suitable for binary classification scenarios and relies on anchor points (samples that clearly belong to a certain class), while the taxpayer industry classification problem has many industry categories and is a multi-classification problem, and the anchor points are difficult to label and get

Images