A 3 Part Python-based project using the libraries pandas, numpy, sklearn to analyze the diabetes dataset from the Medtronic 670G system (artificial pancreas medical control system).
Calculated average per day percentage of records under hyperglycemic and hypoglycemic conditions from the diabetes dataset in the Medtronic 670G system (artificial pancreas medical control system).
Extracted useful meal and no-meal data from over 22,000 data samples recorded by the Continuous Glucose Sensor and Insulin Pump, trained an SVM classifier to categorize into meal and no-meal classes, and achieved an accuracy of 80%.
Implemented K-Means, DBSCAN on generated feature matrix to visualize clusters of similar records with low SSE of 72, 6.