Extracting and cleaning data from semi-structure Chinese texts

Abstract

Data mining helps to uncover valuable information from raw data in large volume. However, the latter usually comes in text instead of structured form, and contains noise which makes analysis difficult. Therefore, it is of vital importance to extract and clean raw data before in-depth analysis are applied. This paper presents a new approach to data extraction and cleaning from semi-structured Chinese texts. Experimental results show that it can effectively prepare data for mining.