Research on Pattern Recognition of English Online Discussion Text Mining Data Based on Corpus

This paper studies a pattern recognition method for English online discussion text mining data based on corpus, aiming to reveal the main language patterns and potential information structures in online discussions. By constructing a special English online discussion corpus, the paper deeply processes a large amount of text data and use text mining technology and pattern recognition algorithms to extract core features. The system design includes three modules: data preprocessing, feature extraction, and pattern classification. Among them, the data preprocessing module mainly completes data cleaning and standardization; the feature extraction module uses word frequency-inverse document frequency (TF-IDF) and word embedding technology (such as Word2Vec) to obtain high-dimensional semantic features; the pattern classification module uses a combination of convolutional neural network (CNN) and long short-term memory network (LSTM) to accurately classify text data. To verify the effectiveness of the algorithm, the paper conducted simulation experiments on the constructed corpus. The results show that the algorithm proposed in this study is superior to traditional methods in both accuracy and recall of pattern recognition. In multiple experiments, the average accuracy of pattern recognition reached 92.3% and the recall rate reached 89.7%, which significantly improved the accuracy of pattern recognition.