This repository contains my work focusing on regression analysis. The task involves exploring the impact of various factors, such as gender, BMI, and blood serum measurements, on blood sugar levels using the diabetes dataset. The assignment explores model fitting, error estimation, and the effect of different learning rates on model performance.
- Blood serum measurement S7 has the most significant impact on blood sugar levels.
- Age and BMI are positively correlated with blood sugar levels, whereas S1, S2, and S3 measurements are inversely related.
- Gender affects blood sugar levels, with males generally having higher levels than females.
- The model shows balanced accuracy for both training and test datasets, indicating a well-fitted model without overfitting or underfitting.
- Experimentation with various learning rates shows a complex relationship between learning rate, training error, and test error, highlighting the importance of selecting an appropriate learning rate for model training.
Sai_Vivek_Rambha_2023585403_Assignment1_Regression_ECS708P.ipynb: The Jupyter notebook containing all code, analysis, and visualizations for the assignment.Sai_Vivek_Rambha_2023585403_Assignment_1_ Regression_ECS708P.pdf: A comprehensive report summarizing the findings, methodology, and conclusions of the regression analysis.
To view or run the notebook, ensure you have Jupyter Notebook installed. You can install Jupyter via Anaconda or using pip:
Clone this repository, navigate to the directory containing the notebook, and start Jupyter Notebook:
Open the .ipynb file to view the analysis.
- Python 3.x
- Jupyter Notebook
- Libraries: numpy, matplotlib, sklearn, etc. (A full list can be found at the beginning of the notebook.)