Machine Learning techniques for Alzheimer prediction

Abstract

From a compilation of 36 different health and lifestyle variants of data for 2500 Alzheimer suspected or diagnosed patients, this project will create a Python code to analyse the data from statistical perspective, and create a model with machine learning algorithms, to obtain the best model for its prediction. Statistical analysis will explore and run some classical tests to data, in function of their nature, determine which variants are more relevant to obtain an Alzheimer diagnostic. Machine Learning techniques will be applied after the statistical results, to obtain the best prediction model from a series of classification algorithms, investigating if it’s better to use the whole dataset available, or our statistical conclusions can lead to simplified and more accurate model to predict successfully a diagnosis. With all the statistical and machine learning results, conclusions are extracted about variants relevancy for diagnosis, and contextualization of the data and the results obtained.