A machine learning supervised analysis using artificial intelligence techniques to predict health outcomes according to risk profiles

The aim is to identify the most predictive exposome components of lung disease and their effect on lung diseases occurrence and severity.

Machine learning is an application of artificial intelligence that provides systems the ability to automatically learn from experience. The process of learning begins with observations of real data in order to look for potential patterns (pattern recognition). Starting from the analysis of a known training dataset, the learning algorithm will produce an inferred function to make predictions about the clinical output values in performing most likely mathematical matching of inputs and outputs, taking into account their statistical variations.

Firstly, conventional techniques such as stepwise logistic regression for classification will be performed as a benchmark to provide useful information to identify single independent predictors.

Secondly, advanced machine learning approaches will be used to fully capture patient’s complexity by determining the best predictors in isolation and combination, and establishing their predictive value on the outcomes of interest.