Machine Learning for substance prediction
Abstract: Implementing a clinical pharmacy support system with a centralized database of drugs and their relationships poses some challenges. One of the challenges is parameterizing drug registers and their attributes, especially when the data originates from different hospitals that use different hospital management systems with multiple standardizations of data sources. It is essential to correlate the associated substance for each drug to identify and alert potential drug interactions. This correlation enables the identification of prescription problems such as drug or therapeutic duplications, cross-reactivity, and y-incompatibility. Manually indicating the substance related to the drug is error-prone and time-consuming, leading to security issues regarding patient care. We aim to develop a model to predict the substance to be associated with a drug based on its description.
Keywords: artificial intelligence, machine learning, substances, clinical pharmacy