Alzheimer’s is one of the chronic diseases that stand as a challenge in the geriatrics domain. The symptoms, diagnosis and treatment varying from person to person makes it more complex to understand the inherent nature of this disease. The lifestyle factors and behavioural traits play major role in the onset and progression of Alzheimer’s disease compared to genetic factors. Studying and analysing such behavioural traits would help the healthcare practitioners to understand its effect on the progression of the disease and its manageability in the advanced stages. This would culminate in enhancing the provision of customized palliative care to alleviate the trauma faced by patients and their caregivers. Making the machines learn from and build models on such data would make the task of healthcare professionals, much easier and quicker. It assists the healthcare stakeholders in studying large amounts of patient data and get equipped to work with patient specific symptoms, to strategize their treatment and improve terminal care facilities. The social media data, available in the form of discussions on patient support groups hosted by Facebook, is manually curated and used for training various machine learning models. Gaussian Naïve Bayes followed by Support Vector Machine were found to be the best performing models based on various context specific evaluation metrics.