Design and Implementation of a Course Answering System Based on Error Correction and Trie Tree Language Model

Abstract

With the rapid growth of network information, the accuracy of input text affects the retrieval results, text proofreading technology emerges as the times require in order to avoid the situation of “wrong answers” caused by wrong questions when searching for information. In the course question answering system, we also need to consider the query speed and efficiency. In order to avoid the trouble of manual proofreading, this paper proposes a system which can automatically correct wrong questions in the professional field of curriculum. Firstly, used the edit distance method for fuzzy matching of error strings; then, used trie tree language model to store data to improve query efficiency. Finally, compared the proofreading effect under different text similarity thresholds, and selected the best value for the experiment. After experimental analysis and comparison, the best result is selected when the text similarity is 0.5, the accuracy rate is 77.91%, the recall rate is 67%, and the F value is 72.04%. Experiments show that the system designed in this paper can effectively correct the wrong text in the field of computer.