Project Title: Detecting Depression from Tweeters Data using LSTM Neural Networks 
Student: Nimrod WYNNE  
Course: MSci Computer Sciene with International Year
Abstract: 
The massive data stream of people’s personal information available on social media 
can be used as a source for data analysis, for example, to help identify mental 
health conditions. The paper analyses the data set collected from Twitter users 
and the notable differences between users who have identified themselves as having 
depression and a control set. This paper describes the methods that entail gathering 
sufficient data, the methods employed to identify depressed users, and pre-processing 
techniques that were implemented. It is demonstrated how the Twitter user base 
can be utilised to create a Long Short-Term Memory (LSTM) network algorithm that 
can identify depressed users on Twitter. Two different approaches were taken: 
classifying whole users’ Tweets versus individual Tweets as either part of a 
depression or a control group. LSTM-based networks appear to be ineffective at 
learning patterns in such large data samples, however, it showed encouraging results 
when applied to individual Tweets.