Support Vector Machine

Support Vector Machine classical method implementation.

This section contains the Support Vector Machine (SVM) implementation for classification tasks. SVM is a supervised learning algorithm that finds a hyperplane to separate data points of different classes with maximum margin.

class TALENT.model.classical_methods.svm.SvmMethod(args, is_regression)

Bases: classical_methods

construct_model(model_config=None)
fit(data, info, train=True, config=None)
metric(predictions, labels, y_info)
predict(data, info, model_name)
class TALENT.model.classical_methods.svm.SVMMethod

Support Vector Machine method for classification tasks.

Key Features:

  • Uses sklearn’s SVC for classification

  • Finds optimal hyperplane for class separation

  • Supports both binary and multiclass classification

  • Automatically handles data preprocessing including normalization and encoding

  • Saves trained model to pickle file for later use

  • Provides probability predictions

Algorithm:

SVM is a supervised learning algorithm that finds a hyperplane to separate data points of different classes with maximum margin. It can handle both linear and non-linear classification using kernel functions.

__init__(args, is_regression)

Initialize the SVM method.

Parameters:

  • args (object) – Configuration arguments containing model settings

  • is_regression (bool) – Whether the task is regression (True) or classification (False)

construct_model(model_config=None)

Construct the SVM model instance.

Parameters:

  • model_config (dict, optional) – Model configuration parameters for SVM

Model Creation:

  • Creates SVC classifier

  • Configures parameters like kernel, C, gamma, etc.

fit(data, info, train=True, config=None)

Train the SVM model on the provided data.

Parameters:

  • data (tuple) – Tuple containing (N, C, y) where N is numerical features, C is categorical features, y is labels

  • info (dict) – Dataset information

  • train (bool, default=True) – Whether to train the model or just load from checkpoint

  • config (dict, optional) – Additional configuration parameters

Returns:

  • time_cost (float) – Training time in seconds

Training Process:

  1. Data Preprocessing: Handles missing values, categorical encoding, normalization

  2. Model Training: Fits the SVM model with optimal hyperplane

  3. Model Saving: Saves the trained model to disk for later use

predict(data, info, model_name)

Make predictions using the trained SVM model.

Parameters:

  • data (tuple) – Tuple containing (N, C, y) where N is numerical features, C is categorical features, y is labels

  • info (dict) – Dataset information

  • model_name (str) – Name of the model for saving/loading

Returns:

  • test_logit (array-like) – Test predictions (probabilities for classification)

Prediction Process:

  1. Data Preprocessing: Applies same preprocessing as training data

  2. Model Loading: Loads the trained SVM model

  3. Prediction: Generates probability predictions

  4. Output: Returns probabilities for classification

Evaluation Metrics:

  • For classification: returns Accuracy, Avg_Precision, Avg_Recall, F1 metrics

References:

[1] Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), 273-297.