class Cadmium::Classifier::Tabular::LogisticRegression

• Cadmium::Classifier::Tabular::LogisticRegression
• Reference
• Object

Overview

Logistic Regression classifier for binary classification.

Uses gradient descent to learn weights that minimize binary cross-entropy loss. Suitable for binary classification tasks with numerical features.

Features

• Fast prediction after training (O(1))
• Probabilistic output
• Works well with large datasets

Example

classifier = Cadmium::Classifier::Tabular::LogisticRegression.new(learning_rate: 0.01, max_iterations: 1000)

features = [
  [1.0, 2.0, 3.0],
  [1.1, 2.1, 3.1],
  [5.0, 6.0, 7.0],
]
labels = ["class_a", "class_a", "class_b"]

classifier.train(features, labels)

result = classifier.classify([1.05, 2.05, 3.05])
# => "class_a"

probs = classifier.classify_probabilities([1.05, 2.05, 3.05])
# => {"class_a" => 0.85, "class_b" => 0.15}

Defined in:

Constructors

• .load_model(path : String) : self

  Load a trained model from a file.

• .new(learning_rate : Float64 = 0.01, max_iterations : Int32 = 1000)

Instance Method Summary

• #bias : Float64

  Get the learned bias term.

• #classify(features : Array(Float64)) : String

  Classify a new sample and return the predicted label.

• #classify_batch(features_batch : Array(Array(Float64))) : Array(String)

  Classify multiple samples at once.

• #classify_probabilities(features : Array(Float64)) : Hash(String, Float64)

  Classify a new sample and return probability scores for both classes.

• #learning_rate : Float64

  Learning rate for gradient descent

• #max_iterations : Int32

  Maximum number of training iterations

• #save_model(path : String) : Nil

  Save the trained model to a file.

• #train(features : Array(Array(Float64)), labels : Array(String)) : self

  Train the classifier using gradient descent.

• #weights : Array(Float64)

  Get the learned weights (for inspection/debugging).

Constructor Detail

classifier = Cadmium::Classifier::Tabular::LogisticRegression.load_model("logistic_regression_model.msgpack")

Instance Method Detail

classifier.classify([1.0, 2.0, 3.0]) # => "class_a"

results = classifier.classify_batch([[1.0, 2.0], [3.0, 4.0]])
# => ["class_a", "class_b"]

probs = classifier.classify_probabilities([1.0, 2.0, 3.0])
# => {"class_a" => 0.85, "class_b" => 0.15}

classifier.save_model("logistic_regression_model.msgpack")

features = [[1.0, 2.0], [3.0, 4.0]]
labels = ["a", "b"]
classifier.train(features, labels)