In the current era in which we live, there is a clear and irreversible tendency to generate and store large volumes of information, from various sources such as: government agencies, public and private companies, clinics and hospitals, social networks, etc.
Hence the great need to analyze the data in order to obtain some benefit for their owner, a third party or humanity in general. With this in mind, we conducted a descriptive and predictive analysis of public medical data of South Africa on patients with possible risk of presenting coronary heart disease (CHD), and applying advanced techniques of supervised machine learning and models calibration, we were able to determine when a person has high probabilities (close to 70%) of presenting or developing this disease, with the objective of being able to contribute to an early detection and diagnosis of it, for further treatment.
Hopeful and convincing results were obtained, which can be improved if there is a greater amount of source data from which to learn.
Click here to see the project and the repository.