Support Vector Machine

IT 위키

Support Vector Machine (SVM) is a powerful supervised machine learning algorithm used for both classification and regression tasks, though it is primarily used in classification. SVM works by finding the optimal boundary, or hyperplane, that best separates the data points of different classes. SVM is effective in high-dimensional spaces and is especially suitable for binary classification problems.

How It Works[편집 | 원본 편집]

SVM aims to maximize the margin between data points of different classes, where the margin is defined as the distance between the closest data points (support vectors) of each class and the hyperplane. By maximizing this margin, SVM achieves a decision boundary that is robust and generalizes well to new data.

If data is not linearly separable, SVM can use a kernel trick to transform the data into a higher-dimensional space, where a separating hyperplane can be found. Common kernels include:

  • Linear Kernel: Suitable for linearly separable data.
  • Polynomial Kernel: Used for non-linear relationships, where the degree of the polynomial can be adjusted.
  • Radial Basis Function (RBF) Kernel: A popular choice for highly non-linear data.
  • Sigmoid Kernel: Sometimes used as an alternative to RBF, though less common.

Applications of SVM[편집 | 원본 편집]

SVM is widely applied across industries, especially in applications requiring high accuracy and interpretability. Common use cases include:

  • Image Classification: Object and facial recognition, where SVM effectively separates complex visual features.
  • Text Classification: Spam filtering and sentiment analysis, where SVM performs well with high-dimensional data such as text features.
  • Medical Diagnosis: Disease prediction and classification of medical data, leveraging SVM’s robustness with noisy and complex datasets.
  • Bioinformatics: Gene expression data classification, where the large number of features and small sample sizes benefit from SVM’s margin-maximizing approach.

Key Parameters in SVM[편집 | 원본 편집]

Several important parameters in SVM influence its performance:

  • C (Regularization Parameter): Controls the trade-off between maximizing the margin and minimizing classification error. A smaller C allows for a larger margin at the cost of more misclassified points, while a larger C aims for accurate classification at the risk of a smaller margin.
  • Gamma (γ): Specific to RBF and polynomial kernels, gamma defines the influence of individual data points. A higher gamma focuses more on close points, potentially capturing complex patterns but risking overfitting.
  • Kernel Choice: Choosing an appropriate kernel is essential, as it directly affects the separation boundary in non-linear problems.

Advantages and Disadvantages of SVM[편집 | 원본 편집]

Advantages:

  • Effective in High-Dimensional Spaces: SVM performs well with high-dimensional data, especially when the number of features exceeds the number of samples.
  • Robust with Clear Margins: Provides strong predictive power with a distinct margin, making it effective for binary classification.

Disadvantages:

  • Computationally Intensive: SVM can be slow to train with large datasets, especially with complex kernels.
  • Less Effective on Noisy Data: Sensitive to overlapping classes or mislabeled data, which can reduce its classification accuracy.

Evaluation Metrics for SVM[편집 | 원본 편집]

To evaluate SVM model performance, several metrics are commonly used:

  • Accuracy: The proportion of correct predictions over total predictions.
  • Precision: The ratio of true positives to all predicted positives, important in scenarios with imbalanced data.
  • Recall: The ratio of true positives to all actual positives, critical in applications where missing positive cases has high costs.
  • F1 Score: The harmonic mean of precision and recall, providing a balanced metric for imbalanced classes.
  • ROC-AUC: The area under the ROC curve, which measures the model’s ability to distinguish between classes across thresholds.

See Also[편집 | 원본 편집]