VisionInsight

Objective

The objective of this project was to develop a machine learning model capable of predicting whether a person has diabetes based on a set of medical variables.

Project structure

1. Data collection
2. Data preprocessing and cleaning
3. Exploratoy data analysis (EDA)
4. Modeling and evaluation
   • Models
     • Logistic Regression
     • K-Nearest Neighbors
     • Support Vector Machine
   • Evaluation
     • F1-Score

Question

1. Is it possible to accurately predict whether a patient has diabetes using diagnostic variables such as number of pregnancies, BMI, insulin levels and age?

Findings

• Random Forest emerged as the best-performing model, achieving an F1-Score of 0.7234.
• The EDA revealed no strong correlation between the number of pregnancies and diabetes outcome.
• Glucose levels and BMI had a stronger relationship with the target variable.

Dataset

UCI Machine Learning & Collaborator. (n.d.). Pima Indians Diabetes Database. Kaggle. https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database

Author

José Habacuc Soto Hernández - SWE Student

• GitHub: https://github.com/habacucsoto
• Portfolio: https://habacuc.dev