Medical transcriptions are widespread and essential in the field of healthcare and its procedures. Patients and staff have to spend a significant amount of manual and mental labour towards these routine tasks and processes. Among multiple endeavours in the automation of tasks in the medical industry, a significant advancement can be done with regard to medical transcriptions. These digitized text-based versions of processed voice reports by healthcare staff, which are formalized reports of singular or multiple encounters with professionals, contain rudimentary and important information about patients, through which there can be a proper data-based mechanism to allow for accurately predicting the specific type of doctor that would suit the patient’s needs. In this category, we are proposing machine learning methods to classify medical transcription content for predicting medical specialties from the already available documents. The highest accuracy received after processing the corresponding documents is via Logistic Regression providing a test accuracy of 83.3%.