Extracting information from texts generated by users of social media platforms becomes a crucial task. In this paper, we describe the systems submitted to the SMM4H shared tasks 1 and 2. The aims of these two tasks are binary and multi-class classification of English tweets. We developed a machine learning-based model integrated with TF-IDF as a feature extraction approach. Four classification algorithms have been implemented namely, support vector machines, passive-aggressive classifier, multi-layer perceptron and random forest. For task 1, the passive-aggressive classifier reported f1-score of 63.7%. For task 2, multi-layer perceptron reported f1-score of 71.4%.