Damilola Alaba HALLY2024-05-212024-05-212022-12Kate TurabianM.Schttps://repository.lcu.edu.ng/handle/123456789/217Depression is a serious mental illness that affects an individual’s professional and personal life. With the development of internet usage people have started to share their experiences and challenges with mental disorder through online platforms. Social media platforms come close to being a true digitization of the human social experience. In many cases people would prefer to express themselves online rather than offline especially in completed suicide attempts around the world. This thesis objectives are to extract despondency indicative social media posts, categorize these posts and then apply an integration of machine learning techniques to generate markers in identifying depressive comments in social media. This will be examined using five algorithms; Support Vector Machine, Logistic Regression, K-Nearest Neighbor, Naves Bayes and Linear Regression. The Analytical hierarchical Performance was used to determine the best algorithm to detect depression in terms of performance metrics. This process identifies users who are at risk of depression to initiate quick intervention. The result of this thesis shows that the support Vector machine has the highest performance metrics. This signifies that the support vector Machine is the best algorithm to apply for the extraction of Despondency symptoms in social media, this will also determine the best machine learning Algorithm that has the best performance and accuracy in detecting despondency symptoms on social media. It is recommended based on the conclusions of this thesis that the support vector machine Algorithm be used in institutions social media platforms to adequately monitor despondency in individuals Keywords: Despondency, Machine learning Algorithm, Analytical Hierarchical Performance, Support Vector Machine, Logistic Regression, Word count: 246.enThesis