This project aims to develop predictive models to assess the risk of diabetes based on various factors such as age, education, income, and health behaviors.
- Early Diagnosis: Identify individuals at risk for early intervention.
- Resource Allocation: Efficiently allocate healthcare resources.
- Public Health: Aid in planning public health campaigns.
Derived from the CDC's BRFSS 2015 survey, the dataset includes 21 feature variables related to health behaviors, chronic conditions, and preventive services utilization.
- Logistic Regression: Serves as a baseline for comparison.
- Decision Trees: Captures complex relationships.
- Naive Bayes: Offers a probabilistic approach.
- Random Forest: Provides high predictive performance.
The models are evaluated based on their ability to predict diabetes status accurately, with a focus on recall to minimize false negatives.
The analysis provides insights into diabetes risk factors and the effectiveness of different predictive models.
- Clone the repository.
- Install required dependencies.
- Run the Python notebook for detailed analysis.
This project is licensed under the IBM License - see the LICENSE for details.