Evaluación del desempeño computacional de sistemas de recomendación aplicado a bases de datos farmacológicas: implementación de un sistema de recomendación híbrido para predicción de interacciones farmacológicas.

Abstract

One of the most serious problems affecting human society is the fear of disease and the health complications that result from it. For this reason, human beings and society have developed what we know as a drug to fight against these fears. Thus, the processes of development and research of possible chemical structures, which could be used safely in humans, has been a very complex research work, given that technological advances have been the basis for the modernization of experimental and analytical methods to improve the properties or active ingredients of drugs. One example of this is the development of computational models, such as recommendation systems, which are responsible for providing the recommendation of an item to a user based on historical data of interactions that relate them. The main challenge that this work aims to address is in making use of recommender systems as a possible approach in the solution to drug repositioning, the search for new interactions for already established chemical compounds. For this reason, this work focuses its efforts on demonstrating that the problem of drug repositioning can be solved at the computational level through a recommendation model, which takes its data flow from historical data of clinical trials performed for compounds and cells of various origins. Carried out experiments gave us a basis of results with acceptable precision, which indicates the feasibility of the application of recommender systems as an alternative to be taken as a starting point in future research focused on drug repositioning