Predicting Diabetes Using Machine Learning

Diabetes could create a global health care catastrophe. The International Diabetes Federation, predicted that there are currently 382 million diabetics worldwide and this figure will treble by 2035 to 592 Million Diabetics! Diabetes is increased blood glucose levels are a characteristic. Diabetes is one of the main risk factors for renal failure, heart failure, and stroke., amputations, blindness and kidney failure and its symptoms are frequent urination, increased thirst and appetite. Diabetes also plays a significant role in heart attacks, strokes, heart failure and blindness among many other complications that affect quality of life such as blindness. As we consume food, our bodies convert it to glucose for use as fuel, which then triggers insulin release from our pancreas. [13] Insulin acts like a key that unlocks cells to allow glucose in and fuel us -however diabetes makes this process ineffective - it affects one or both types of individuals in society. There are both typical forms and unusual variations of diabetes, such as gestational diabetes. [6] Machine Learning, which explores how computers learn through experience, was chosen as the focus for this work in order to develop an early diabetes prediction system that more accurately meets patient needs by combining several machine learning methodologies such as K Closest Neighbor Algorithm (KCLNA), Logistic Regression, Decision Tree Support Vector Machine and Random Forest; then compute accuracy scores for each model before selecting those with good accuracy scores for use when making predictions about gestational diabetes diagnosis.