A system and method for providing encoding for target data

Abstract

The invention discloses a system and a method for providing codes for target data, and belongs to the technical field of tax data application. The system comprises a training module which obtains training data, the training data comprises a classification coding table and historical invoice data, the classification coding table and the historical invoice data are trained, a training result is obtained, and a plurality of training models are generated based on the training result; The model merging module is used for merging the plurality of groups of training models based on the plurality of groups of training models and superposing the same data set of the training result; And the code providing module reads the training result data, is provided with a plurality of interfaces, receives the target data needing to be coded and transmitted by the interfaces, and provides coding information for the target data needing to be coded. Historical data is subjected to detailed and effective preprocessing according to actual data conditions, interference information is removed, and the training accuracy is improved.

Description

technical field [0001] The present invention relates to the technical field of tax data application, and more specifically, to a system and method for providing coding for target data. Background technique [0002] The latest "Coding Table of Goods and Services Tax Classification" released by the State Administration of Taxation strictly divides goods and services into 4207 categories, including 675 major categories and 3532 small categories. In 2016, the National Taxation Bureau issued a request to start trial adding tax classification codes and coding-related functions to the invoicing software. In addition, there are a large number of errors and inaccurate manual marking codes in the invoices issued by the local end, which have a misleading effect on the statistics and analysis of preventing corporate tax evasion and tax evasion based on the corporate tax rate and input and sales analysis of the commodities sold. Due to the limitations of professional knowledge and energ...

AI Technical Summary

Problems solved by technology

However, because Naive Bayes is a statistical algorithm and has certain limitations based on the independent assumption of feature conditions, first of all, it cannot handle unregistered words, and less data in the sample or too many misclassified data in the sample will lead to inaccurate classification. Accurate, consistent processing of word segmentation in different positions in the product name, each recognition needs to calculate the probability of all categories, the calculation cost is high and the communication bandwidth occupies a large amount, making it impossible to actually use it online, and an optimization plan is urgently needed

[0004] Traditional database systems store all data on disks, which not only requires frequent access to disks for data read operations, but also lowers performance when the amount of data is large and read operations are frequent

Images