The project detailed here attempts to build a machine learning model that will help classify individuals to whether they have diebetes or not,then further recommend foods to both the classified classes independently.
According to the National Institute of Diabetes and Kidney Diseases, diabetes is a disease that occurs when your blood glucose (blood sugar) is too high. Blood glucose is your main source of energy and comes from the food we eat. Insulin, a hormone made by the pancreas, helps glucose from food get into your cells to be used for energy. Sometimes the body does not make enough or any insulin or doesn’t use insulin well. Glucose then stays in your blood and doesn’t reach your cells. Food is among the most basic needs for the control of diabetes. Proper dietary habits can generally promote good health. However, each person’s diet needs are typically varied based on some individual factors including gender, age, and physical differences or health status. In addition, each person usually differs in terms of food preference. Thus a diet selection that balances between the individual need and preference is often challenging. In this regard, a food nutrition recommendation system would be helpful. A food nutrition and recommendation system is a model which suggests the best diets according to a patient's health situation and preferences. The system recommends food based on the users’ age, blood glucose, blood pressure, BMI, insulin and diabetes pedigree function.
Hypothesis: Classify individuals into two classes.Those with diabetes and those without
X-Variables: The feature variables used to predict the outcome.
y-Variables: The diabetes class labels (0, 1) (without,with respectively)
Experimental setup: Classify the individuals to those with diabetes (1) and those without(0) based on the individuals prompted inputs of their medical details.
Design of the Experiment: Analyze 768 individuals over 9 metrics use to predict their outcome class then recommend food for each individual class independently.
Sample size: 768 individuals
Then to use the nutrition dataset to recommend diets based on the glycemic index of food chosen by the respective individuals
The data we used for this analysis was from the PIMA people of indian heritage. Their information were taken as part of a clinical trial therefore its accuracy is of the highest percentile, thorough and verifiably relevant in every accord.
Diabetes dataset https://drive.google.com/file/d/1pJ_OfcokEAQfyejGntU7v7FJ8HQ1joJG/view?usp=sharing
Glycemic index Dataset https://docs.google.com/spreadsheets/d/1BwRHgNKghr4H0ZtpFvrG4S6WEk94JQFt/edit?usp=sharing&ouid=109390265475617060608&rtpof=true&sd=true
Distributed under the MIT License. See https://github.com/Eltonjohn-Oketch/EURISKO-Food-Nutrition-And-Dietetics-for-Diebetics-/blob/main/LICENSE for more information.
Hypertuned XGBoost
Support Vector Machine
Neural Networks (MLP
XGBoost
Random Forest
K-Nearest Neighbors
Decision Trees
Hypertuned KNN
MLP classifier
Sequential layers.
Eurisko - (Eltonjohn) scrum master