Multi-Disease Prediction with Artificial Intelligence from Core Health Parameters Measured through Non-invasive Technique

Healthcare falls under the basic amenities to be provided to the society. In the current world, Non-invasive diagnostic tests are of prime requirement in the medical field to avoid the painful process of collecting samples. This work aims to develop a portable and affordable diagnostic system to measure basic health parameters such as Body mass index (BMI), temperature, pulse rate, blood oxygen saturation level, blood pressure, and blood glucose content and use the machine learning (ML) models to predict the future health complications that could be used to take preventive measures. In this work, we demonstrated that using only core health parameters many diseases can be predicted with good accuracy. Diabetes is taken as a case study and predicted using the Pima Indian Diabetes Dataset (PIDD) of UCI Machine learning Repository. Diabetes is predicted by extracting only core parameter data and compared with the prediction using the complete data set and resulted in comparatively similar accuracy. Decision tree, Naive Bayes, and K-neighbors classifiers (KNN) are used. KNN outperforms the other two algorithms in prediction in both data sets.