基于MongoDB的文本分类研究

Research on text classification based on MongoDB

文章基于流行的非关系型数据库MongoDB,结合Spark机器学习库中的朴素贝叶斯分类器和支持向量机,对豆瓣影评及京东商评进行情感分类,并采用准确率、召回率、F-Measure等指标对分类效果进行评价,最后测试了Spark-MongoDB平台的扩展性能。

Based on the popular non-relational database： MongoDB, this paper combines naive Bayesian classifier and support vector machine in Spark machine learning library to do emotion classification for Douban film review and Jingdong commercial review, and takes advantages of accuracy, recall rate, F-Measure and other index to make evaluation of the classification effect, and finally test the extended performance of Spark-MongoDB platform.