WeiboRank (English Version)

Weibo (a Microblog site popular in China) User Sentimental Influence Analysis Results

Results site

1. Collab Image
2. User Ranklist

• Analysis text contents of Weibo and extract sentimental infomation with Python.
• Define the weight of the followed and following link relationship with Python.
• Classify and rank sentimental influence of Weibo users among the SNS using PageRank iteration process.
• Visulize sentimental analysis results of ranking and relationship network of users with Matplotlib and D3.js.

Sentimental Analysis

Use Word2Vec to do the Word Embedding to represent words. Use Recurrent Neural Network(RNN)/LSTM (tutorial: keras) to train the sentimental classification.

Steps:

1. get data and input data into matrix
2. Use jieba to cut the Chinese sentence into words
3. Word2Vec model set up, training, testing, fine-tune, training, testing, fine-tune...... until -> :)
4. Get the three types of sentiments: positive, negative and neutral of these sentences.
   

Network Structure

A subdomain/second-level domain for project github pages

Final target is: Second-level domain: https://jeness.github.io/WeiboRankEngVer/CollabEngVer/ -> weiborankengver.haoranyu.info/CollabEngVer main page domain: 'jeness.github.io' -> www.haoranyu.info

Step 1: In gh-pages branch in WeiboRankEngVer repo, add CNAME file, add weiborankengver.haoranyu.info in the CNAME file.

Step 2: add a new CNAME in godday

http://zangbo.me/2017/06/12/GithubPageDomainName/