This is a course project of "Computational Methods and Tools for Data Science, National Taiwan University, 2016 Spring"

In this project we aim to build an intelligent system that can chat with a person with emotion considered. You may view our project description slides here.

In the short term, we aim to do some in depth analysis of how people chat with others on social media and messanger. This can allow more sophisticated design of our final system.

In milestone2, we focus on collecting and analysising data. We release four tools:

• data_crawling: Crawl data from Facebook chatroom log & DCard website.
• tf-idf: Tf-idf analysis to extract keywords from each chatroom for further analysis.
• visual_LM: Building language model for each chatroom and try to do some analysis (including nearest neighbor search, multidimensional scaling analysis).
• word2vec: Using neural network non-linear model to embed each word into a meaningful vector space for further analysis.

Go to Data_analysis/ for more detailed description for each module.

蔡捷恩(@tsaiJN), [email protected]

黃兆緯(@whsunset), [email protected]

李承軒(@Wariard), [email protected]

朱瑤章(@dswf65411), [email protected]

This project is released under the BSD 2-Clause License