This is mainly for codes of nlp tasks.
实现了最简单的前后向最大匹配算法。
实现了CYK方法,主要参考了http://f.dataguru.cn/thread-693052-1-1.html。
使用PySpark实现,特征选择的方法参考了论文《Thumbs up? Sentiment Classification using Machine Learning Techniques》by Bo Pang, Lillian Lee and Shivakumar Vaithyanathan。
实现了基于特征的单文本摘要,和基于TextRank的文本摘要技术。