SVM-Project

SVM (Use of SVM Kernels on formed and real data using object-oriented programming)

PROJECT 1:

1. Formed Data using uniform distribution (HW4.py)

2. Defined hyperplane by polynomial with degree 4 and used that as separating hyperplane (HW4.py)

3. Standardized Data (HW4.py)

4. Split data into test and train sets (HW4.py)

5. Linear SVM (HW4.py / SVM.py)

• Tune the parameter Cost (C) (HW4.py / SVM.py)
• Fit model with best parameters (HW4.py / SVM.py)
• Look at results of model via Confusion Matrix, Histogram Plots, and Scatter Plots (HW4.py / SVM.py)

6. Radial Kernel (HW4.py / SVM.py)

• Tune the parameters Cost and Gamma (HW4.py / SVM.py)
• Fit model with best parameters (HW4.py / SVM.py)
• Look at results of model via Confusion Matrix, Histogram Plots, and Scatter Plots (HW4.py / SVM.py)

7. Polynomial Kernel (HW4.py / SVM.py)

• Tune the parameters Cost, Coef0, and degree (HW4.py / SVM.py)
• Fit model with best parameters (HW4.py / SVM.py)
• Look at results of model via Confusion Matrix, Histogram Plots, and Scatter Plots (HW4.py / SVM.py)

8. PDF Describing entire process and showing the formation of data, data description, tables, and graphs (Math6350.HW4.Vasquez.pdf)

PROJECT 2:

1. Exploration of Data (explore.py)

2. Feature Engineering (explore.py)

3. Reduction of Cases for project guidelines (explore.py)

4. Creating Classes for linear Regression Problem (total of 3 classes)

5. Standardize (HW4part2_proportion_2.py / test_train.py)

6. Test Train Split (HW4part2_proportion_2.py / test_train.py)

7. Radial Kernel (HW4part2_proportion_2.py / SVM.py)

• Tune the parameters Cost and Gamma (HW4part2_proportion_2.py / SVM.py)
• Fit model with best parameters (HW4part2_proportion_2.py / SVM.py)
• Look at results of model via Confusion Matrix, Histogram Plots, and Scatter Plots (HW4part2_proportion_2.py / SVM.py)
• Terminal Predictions using CDF from each SVM created for each class 1 vs all, 2 vs all, and 3 vs all (HW4part2_proportion_2.py / SVM.py)
• Terminal Confusion Matrix and analysis of model (HW4part2_proportion_2.py)

8. Polynomial Kernel (HW4part2_proportion_poly2.py / SVM.py)

• Tune the parameters Cost and degree (HW4part2_proportion_poly2.py / SVM.py)
• Fit model with best parameters (HW4part2_proportion_poly2.py / SVM.py)
• Look at results of model via Confusion Matrix, Histogram Plots, and Scatter Plots (HW4part2_proportion_poly2.py / SVM.py)
• Terminal Predictions using CDF from each SVM created for each class 1 vs all, 2 vs all, and 3 vs all (HW4part2_proportion_poly2.py / SVM.py)
• Terminal Confusion Matrix and analysis of model (HW4part2_proportion_poly2.py)

9. PDF Describing entire process and showing the formation of data, data description, tables, and graphs (Math6350.HW4part2.Vasquez.pdf)