Chinese address word segmentation tagging method

Abstract

The invention relates to a Chinese address word segmentation tagging method. The method includes: step 11, artificial word segmentation and labeling selected address data as training data; step 12, for single Arabic numerals or English alphabetic characters and multiple continuous Arabic numerals or English alphabetic characters, specifying single Arabic numerals or English alphabetic characters Numerical characters or English alphabetic characters are replaced; step 13, the training data is converted into the format required by the CRF++ tool; step 14, the feature template is defined; step 15, the word segmentation model and the labeling model are respectively established using the CRF++ tool; step 16, a single that appears in the address Arabic numerals or English alphabetic characters and multiple consecutive Arabic numerals or English alphabetic characters are replaced by the specified single Arabic numerals or English alphabetic characters; step 17, word segmentation with CRF++ tool; step 18, restore the Arabic before replacement Numeric characters or English alphabetic characters. The Chinese address participle labeling method of the present invention has a high accuracy rate.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a Chinese address word segmentation tagging method. Background technique [0002] When people fill in the recipient address, office address, home address and other address information, they usually write together the province, city, district, house number, residential area, room number and other information to form address information, such as "Guangdong Shenzhen Nanshan District 713, Software Building, Science and Technology Park, No. 9, High-tech Middle Road, this writing method is suitable for manual recognition, and when the structural details of the address are recognized by the machine, the first processing is word segmentation and labeling, and the input long text is divided into one by one words, and mark the attributes of the words. For example, the word segmentation result of the address information in the above example may be: "Guangdong / Shenzhen, Nanshan Distr...

AI Technical Summary

Problems solved by technology

However, the accuracy rate of Chinese address word segmentation and tagging by applying general word segmentation tagging or entity recognition technology is not very high, only about 80%.

Images