Corpus Based Chinese Grammar Error Detection Rules Evaluation Method and System

Accurate detecting and correcting Chinese grammar errors is the problem that many Chinese NLP research experts have been working on. However, Chinese grammars are complex and flexible. Collecting and Filtering the Chinese grammar errors with good popularity has been a major problem in this area. This paper presents an indexing method of Chinese grammar rules based on Corpus. With the corresponding searching system, we can evaluate the rules of the frequency and the accuracy. For each evaluated rule, we adopt an iterative approach to improve it for its better performance in the real world. Meanwhile, it also introduces the important role of corpus of this method. The test results on different corpus show that the feasibility and effectiveness of this method.