COMP_9800_IEEE_Rutvij_Vardhil.pdf (1.15 MB)
Download file

Depression sentiment analysis based on social media content like Twitter

Download (1.15 MB)
posted on 2022-12-11, 16:15 authored by Rutvij KananiRutvij Kanani, Jinan FiaidhiJinan Fiaidhi, Vardhil PatelVardhil Patel

Everyone may now readily communicate and share their sentiments with individuals around the world thanks to the various social networking platforms. In the current scenario, some people are more extroverted on social media than others around them. In the end, whenever they get into trouble or any situation where they need other people’s help who gives them motivation or who cares, they would not be there because that time sufferer prefers to express their feeling on social media rather than any close once. Therefore, if they are warned in advance, there are some strategies that reduce the stress and mental  health issues that they are experiencing. Due to rising these issues globally, attract many researchers to  focus on the subject and provide some viable solutions, where there is still a need for more research that  provides some efficient results. So finally, we built a project in which it takes users’ tweets as input, and  it will give the result in the form of depression or not using the help of a machine learning algorithm. For  the segmentation of tweets, we are using LSTM (Long Short-Term Memory) machine learning algorithm  which is best suited for this task. Not only this but LSTM performs best out of available machine learning  algorithms. All the emotions are identified as neutral, positive, or negative which assists to provide a  solution toward sentiments. Apart from this, we are not focusing on only the English language but we  will try to counter this issue in different languages by translating other languages into English, which  other researchers have missed out on in their research. With the help of these results, the company will  take different steps to mitigate the stress from those users by providing motivational feeds in their feed  section or any other way. 


Email Address of Submitting Author

Submitting Author's Institution

Lakehead University

Submitting Author's Country

  • Canada