qyfang/TextClassification

Implements a complete text classification pipeline using jieba tokenization, TF-IDF weighting, and parallel processing via process pools to accelerate both web scraping and preprocessing across the 1-million-document dataset. LinearSVC with grid search and 5-fold cross-validation achieves 0.90 average F1-score, significantly outperforming the Naive Bayes baseline (0.79), with visualization tools comparing classifier performance via confusion matrices and performance histograms.

110 stars. No commits in the last 6 months.