multi-class-svm

Implemented a multi-class support vector machine with kernel function tricks by solving the dual quadratic programming problem of SVM.

Additional Information:

1. Implemented an interface for selecting the type of kernel function.
2. The demo uses the Iris flower dataset and can be executed directly with the command python nlsvm.py.
3. zip.train and zip.test are the training samples and test samples of the handwritten digit dataset, respectively.
4. The quadratic programming solver used is osqp
5. The testing environment is python 3.9.7, and it requires a minimum ofpython 3.9
6. The open-source license used is MIT License

example：

if __name__ == '__main__':
    from datetime import datetime
    import os
    os.chdir(os.path.dirname(__file__))
    from sklearn.model_selection import train_test_split
    from sklearn import datasets

    starttime = datetime.now()
    iris_df = datasets.load_iris()  # Load the Iris dataset
    X = iris_df.data
    y = iris_df.target.astype(int).astype(str)
    
    # Split the dataset into a training set and a test set
    X_train, X_test, y_train, y_test = train_test_split(
        X, y, test_size=0.4, random_state=42, stratify=y)
    
    print("Training on the Iris dataset...")
    # Use the Gaussian kernel function with sigma=1.0
    model = nlsvm(X_train, y_train, cost=10.0, kernel="rbf", sigma=1.0)
    
    print("Testing the classifier...")
    y_pred = predict(model, X_test)
    print(f"Test set accuracy: {accuracy(y_pred, y_test)*100:.2f}%")
    endtime = datetime.now()
    print("Total time: ", (endtime - starttime).seconds, "seconds")

output:

Training on the Iris dataset...
classes:  ['0' '1' '2']
A total of 3 classifiers are required for this task...
Training classifier: 1/3 ['0', '1']
Training classifier: 2/3 ['0', '2']
Training classifier: 3/3 ['1', '2']
Training complete!
Number of training samples: 90; Dimension: 4; Number of classes: 3
Testing the classifier...
Number of test samples: 60
Test set accuracy: 98.33%
Total time:  0 seconds