Description

Analysing suicidal tweets using machine learning involves using algorithms and natural language processing techniques to analyze large amounts of text data from social media platforms, with the goal of identifying patterns and signs of suicidal behavior. The process typically involves text pre-processing, feature extraction, and the use of machine learning models such as sentiment analysis, text classification, and risk assessment. The output from these models can provide valuable information for suicide prevention and early intervention efforts by identifying individuals who may be at risk and providing them with the necessary resources and support. The use of machine learning in this field has the potential to revolutionize the way we approach suicide prevention, making it more effective and scalable, and potentially saving many lives.

Suicidal behaviour is a severe public health issue that affects people all over the world.

Appropriate risk assessment by mental health practitioners and timely support is critical to the

efficacy of suicide prevention. However, most people don't receive any treatment due to the

limited available mental health professionals, the lack of understanding of mental health, and the

stigma associated with mental illness. Considering the above facts, it becomes crucial to

identify the suicidal risk/ideation through observation than to rely only on the self-report.

Moreover, traditional methods like psychological battery tests and clinical judgement

can't provide the assessment of suicidal risk in real-time, thus delaying the reporting of

at- risk individuals. Suicidal thoughts are increasingly expressed on online platforms.

If these suicidal posts are recognised early through the intelligent mechanism,

many lives could be saved. This thesis inspects and examines the feasibility of automatically

detecting suicidal content from non-suicidal content on social media and differentiating the

content based upon the severity of the message using natural language processing and a

machine learning approach.