A predictive Analysis System for Feature Optimization using Genetic Algorithm

How to run code

1. Clone the repository

https://github.com/hbkabir004/Feature-Optimization-using-Genetic-Algorithm.git

2. Install Python

2.1 Download the Python Installer binaries.

https://www.python.org/downloads/

2.2 Run the Executable Installer.

2.3 Add Python to environmental variables.

2.4 Verify the Python Installation.

3. Install & Import Necessary Python Packages in Visual Studio Code

Follow the tutorial to install & import necessary python packages in VS Code

4. Open the project in the IDE (VS CODE recommended)

Follow the tutorial to Run Python in Visual Studio Code on Windows 10

5. RESULTs

5.1 RANDOM FOREST

Optimal Feature Set

['radius_mean', 'texture_mean', 'area_mean', 'smoothness_mean', 'compactness_mean', 'concavity_mean', 'concave points_mean', 'symmetry_mean', 'radius_se', 'perimeter_se', 'concavity_se', 'radius_worst', 'texture_worst', 'area_worst', 'concave points_worst', 'fractal_dimension_worst']

Feature Importances

[0.0345544  0.01528677 0.05257772 0.04909357 0.00796349 0.0073864
 0.03402101 0.0742245  0.00459763 0.00389691 0.00996118 0.00515219
 0.01823021 0.04250951 0.00312534 0.00529447 0.00446035 0.00320924
 0.00378835 0.00521066 0.14585278 0.02144762 0.14566475 0.09814828
 0.01441097 0.01654489 0.04011699 0.1168934  0.01104655 0.00532988]

Optimal Accuracy = 99 %

Average Accuracy saved  0.961335676625659
Average Precision       0.9612987777153709
Average Recall          0.9557766502827546
Average F1-Score        0.9584064327485381

     B    M
B  349    8
M   14  198

5.2 Light GBM

Optimal Feature Set

 ['smoothness_mean', 'concavity_mean', 'concave points_mean', 'perimeter_se', 'area_se', 'concave points_se', 'radius_worst', 'texture_worst', 'perimeter_worst', 'area_worst', 'smoothness_worst', 'compactness_worst', 'concavity_worst', 'concave points_worst', 'symmetry_worst', 'fractal_dimension_worst']

Feature Importances

[ 69 170  49  36  65  64  88 174  81  54  80  79  48 142  54  43  27  47
  57  47  92 282 147 144 104  40 109 206 107  42]

Optimal Accuracy = 99 %

Average Accuracy saved  0.9718804920913884
Average Precision       0.9707985143918292
Average Recall          0.9689696633370328
Average F1-Score        0.9698694696708942

     B      M
B   350     7
M   9     203

5.3 XGBoost

Optimal Feature Set

['perimeter_mean', 'smoothness_mean', 'compactness_mean', 'concavity_mean', 'concave points_mean', 'fractal_dimension_mean', 'concavity_se', 'concave points_se', 'symmetry_se', 'fractal_dimension_se', 'texture_worst', 'area_worst', 'compactness_worst', 'concave points_worst', 'symmetry_worst']

Feature Importances

[0.00523759 0.01394322 0. 0.01937084 0.00388636 0.00419483
0.00792941 0.12601759 0.00246526 0.00336443 0.00851927 0.01265129
0.00763103 0.00893879 0.00833281 0.0060065 0.01219247 0.01192982
0.00208353 0.00228187 0.3791942 0.0181876 0.19956343 0.01817427
0.00790597 0.00302455 0.01822192 0.07777867 0.00269732 0.0082751 ]

Optimal Accuracy = 99 %

Average Accuracy saved  0.9648506151142355
Average Precision       0.966116035455278
Average Recall          0.9585777707309339
Average F1-Score        0.9621111229490731

     B      M
B   351     6
M   14    198