On Learning from Natural Language to Develop Intelligent System to Identify Potential Medical Condition
Abstract With the rise in human population and emergence of the medical crisis across the whole globe, it has become essential to develop smart systems to automate the process of identification of medical conditions and therefore provide timely aid to the patient. As the first step to develop such a system in this paper we aim to create a natural language based medical condition identification system. The user would provide a text review of how they feel with few other categorical features, based upon that our model would identify the potential medical condition the user is suffering from. We employed three different machine-learning algorithms and mitigated class imbalance in our dataset. Empirical results indicate that Random Forest is the best machine-learning algorithm among all the investigated models with an accuracy of 80.39%, whereas the accuracy of the AdaBoost model improved the highest with an absolute value of 7.77% after mitigating class imbalance