lda_classifier

Linear Discriminant Analysis classifier for continuous datasets using a shared pooled covariance estimate with diagonal regularization. The implementation learns one linear discriminant function per class and predicts the class with the highest discriminant score.

The library implements the classifier_protocol defined in the classification_protocols library. It provides predicates for learning a classifier from a dataset, using it to make predictions, inspecting class scores, and exporting it as a list of predicate clauses or to a file.

Datasets are represented as objects implementing the dataset_protocol protocol from the classification_protocols library. All dataset attributes must be declared as continuous.

API documentation

Open the ../../docs/library_index.html#lda_classifier link in a web browser.

Loading

To load this library, load the loader.lgt file:

| ?- logtalk_load(lda_classifier(loader)).

Testing

To test this library predicates, load the tester.lgt file:

| ?- logtalk_load(lda_classifier(tester)).

Features

  • Continuous Datasets: Accepts only datasets whose attributes are all declared as continuous.

  • Pooled Covariance Model: Learns a shared covariance matrix and class-specific means and priors.

  • Regularized Estimation: Applies configurable diagonal regularization to stabilize covariance inversion.

  • Feature Scaling: Supports optional z-score scaling of continuous features before fitting the model.

  • Score Inspection: Provides discriminant scores for all classes using predict_scores/3.

  • Classifier Export: Learned classifiers can be exported as predicate clauses or written to a file.

Options

The learn/3 predicate supports these options:

  • feature_scaling/1 - whether to standardize continuous attributes before training (default: true)

  • regularization/1 - positive diagonal value added to the pooled covariance matrix before inversion (default: 1.0e-6)

Usage

Learning a classifier

| ?- lda_classifier::learn(iris_small, Classifier).

| ?- lda_classifier::learn(iris_small, Classifier, [regularization(1.0e-5)]).

Making predictions

| ?- lda_classifier::learn(iris_small, Classifier),
     lda_classifier::predict(Classifier, [sepal_length-5.1, sepal_width-3.5, petal_length-1.4, petal_width-0.2], Class).

| ?- lda_classifier::learn(iris_small, Classifier),
     lda_classifier::predict_scores(Classifier, [sepal_length-6.0, sepal_width-2.9, petal_length-4.5, petal_width-1.5], Scores).

Exporting the classifier

| ?- lda_classifier::learn(iris_small, Classifier),
     lda_classifier::export_to_clauses(iris_small, Classifier, classify, Clauses).

| ?- lda_classifier::learn(iris_small, Classifier),
     lda_classifier::export_to_file(iris_small, Classifier, classify, 'classifier.pl').

Classifier representation

The learned classifier is represented as a compound term with the form:

lda_classifier(Encoders, Models, Options)

Where:

  • Encoders: list of continuous feature encoders with learned scaling parameters

  • Models: list of class_model(Class, Prior, Mean, Weights, Offset) terms

  • Options: merged training options used to learn the classifier

When exported using export_to_clauses/4 or export_to_file/4, this classifier term is serialized directly as the single argument of the generated predicate clause so that the exported model can be loaded and reused as-is.

References

  1. Fisher, R.A. (1936). “The use of multiple measurements in taxonomic problems”.

  2. Hastie, T., Tibshirani, R. and Friedman, J. (2009). “The Elements of Statistical Learning”. Section 4.3.

  3. Bishop, C.M. (2006). “Pattern Recognition and Machine Learning”. Section 4.1.