Machine Learning for Diabetes Prediction
The machine learning input-output process, visualized.Copyright 2018 Francesco Pettini, CC-BY-4.0, Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Machine_learning.jpg
Machine learning is an exciting method that engineers can use to understand large data sets. In this hands-on activity, students put on their computer science hats to tackle a real-world problem: designing a machine learning model that can predict whether a patient has diabetes. Students first learn about the diabetes epidemic and the relationship between machine learning and healthcare. They design a simple program using machine learning that can predict whether a patient has diabetes depending on various symptoms and measurements. The goal is not just to expose students to machine learning, but the realities of the diabetes epidemic.
Engineers use machine learning in various applications across all industries. Recently research studies have explored the benefits of using a machine learning approach to detect anomalies in the area of healthcare. Machine learning engineers continually evolve the technology to improve the performance and the accuracy of the results, such as detecting diabetes cases. Much research has gone into the non-invasive automated detection of diabetes using machine learning techniques. Machine learning is employed based on steps of data preprocessing, feature selection, and classification.
After this activity, students should be able to:
- Participate in an inquiry-based activity where they will make a hypothesis on which columns from a dataset, representing different symptoms to predict if a person has diabetes.
- Use machine learning skills to design a model to make the predictions they hypothesized.
- Use the model to test other symptom data and refine their research.
- Utilize Google Collab and Python to introduce machine learning tools for predicting diabetes.