Regresar

Predicción de Covid19 con el uso del Algoritmo Random Forest y Redes Neuronales Artificiales

Abstract:

Currently SARS-CoV-2 or Covid19 as it is known, has variants or mutations that spread rapidly affecting people, without the health professionals being able to detect it in a timely manner to give an adequate treatment and thus be able to control its spread. This manuscript describes the implementation of an analysis and prediction model of the spread of Covid19, which through artificial intelligence techniques related to Machine Learning, will allow the application of supervised learning strategies to programs developed in the Python programming language. so that when processing large volumes of data they can learn from past experiences and allow new inputs to be processed, generating prediction information quickly and reliably. The approach of making an analysis on a data set extracted from an open-source will serve to later carry out an exploratory analysis of the processed. Three predictions were made, which are: If the patient has SARS-CoV-2, days elapsed until mortality and mortality from covid, using classification and regression algorithms that, according to previous studies, allowed the selection and application of the Random algorithmic model Forest and Artificial Neural Networks whose reliability metrics allow us to accept the expected predictions for an adequate decision making.