直播技术>> Sentiment-Analysis>> 返回
项目作者： akjayant

项目描述 ：
Twitter Sentiment Analysis
高级语言： Python
项目地址: git://github.com/akjayant/Sentiment-Analysis.git
创建时间： 2017-09-04T10:39:34Z
项目社区：https://github.com/akjayant/Sentiment-Analysis
开源协议：
下载


Sentiment-Analysis

  1.  Sentiment analysis on hackerearth predict happiness challenge

accuracy - 88.592%
Model - Ensemble of Random Forest, Gradient Boost, Logistic Regression and Naive Bayes

  1.  Real Time Twitter Sentiment Analysis

So I pickled the models created by hackerearth competition and used it for making predictions in another cool thing -
Twitter Sentiment Analysis but so as to make it look professional & meaningful I made it end to end. 

  1. TWITTER ---tweets--stream----> PYTHON MACHINE LEARNING CODE---sentiment--tagged--tweets---> mySQL DB

Hence as many times I will run it will keep accumulating results and simulataneously store in database can serve for future
modelling as well (I can manually tag the incorrect predictions and make model stronger) . It can stream live tweets thanks to awesome
twitter api and predict sentiment of it! I had pretty good results on #flipkart #bigbillionday sale vs #amazon #greatindainsale which
analysed twitter sentiments of customers.