All Examples
Example 08
beginner
08
Classification
ML Basics

Logistic Regression

Logistic regression models the probability of class membership using a sigmoid function. This example loads the Iris dataset, creates a binary classification target (setosa vs. rest), splits and scales the data, trains a LogisticRegression model, and evaluates with accuracy, precision, recall, F1 score, and confusion matrix.

Deepbox Modules Used

deepbox/datasetsdeepbox/mldeepbox/metricsdeepbox/preprocess

What You Will Learn

  • Train a LogisticRegression classifier for binary classification
  • Preprocess with StandardScaler for better convergence
  • Evaluate with accuracy, precision, recall, F1 score
  • Read confusion matrices: rows are true labels, columns are predictions

Source Code

08-logistic-regression/index.ts
1import { loadIris } from "deepbox/datasets";2import { accuracy, confusionMatrix, f1Score, precision, recall } from "deepbox/metrics";3import { LogisticRegression } from "deepbox/ml";4import { tensor } from "deepbox/ndarray";5import { StandardScaler, trainTestSplit } from "deepbox/preprocess";67console.log("=== Logistic Regression ===\n");89// Load the famous Iris dataset for classification10const iris = loadIris();11console.log(`Dataset: ${iris.data.shape[0]} samples, ${iris.data.shape[1]} features\n`);1213// Simplify to binary classification: setosa (0) vs non-setosa (1)14const y_binary: number[] = [];15// Convert multi-class labels to binary16for (let i = 0; i < iris.target.size; i++) {17  const label = Number(iris.target.data[iris.target.offset + i]);18  // Map setosa to 0 and other classes to 119  y_binary.push(label === 0 ? 0 : 1);20}21const y = tensor(y_binary);2223// Split into train (70%) and test (30%) sets24const [X_train, X_test, y_train, y_test] = trainTestSplit(iris.data, y, {25  testSize: 0.3,26  randomState: 42,27});2829console.log(`Training set: ${X_train.shape[0]} samples`);30console.log(`Test set: ${X_test.shape[0]} samples\n`);3132// Standardize features (mean=0, std=1) for better convergence33const scaler = new StandardScaler();34scaler.fit(X_train);35const X_train_scaled = scaler.transform(X_train);36const X_test_scaled = scaler.transform(X_test);3738console.log("Features scaled\n");3940// Create and train logistic regression classifier41const model = new LogisticRegression({ maxIter: 1000, learningRate: 0.1 });42model.fit(X_train_scaled, y_train);4344console.log("Model trained!\n");4546// Make predictions on test data47const y_pred = model.predict(X_test_scaled);4849// Calculate classification metrics50const acc = accuracy(y_test, y_pred);51const prec = precision(y_test, y_pred);52const rec = recall(y_test, y_pred);53const f1 = f1Score(y_test, y_pred);5455console.log("Model Performance:");56console.log(`Accuracy:  ${(Number(acc) * 100).toFixed(2)}%`);57console.log(`Precision: ${(Number(prec) * 100).toFixed(2)}%`);58console.log(`Recall:    ${(Number(rec) * 100).toFixed(2)}%`);59console.log(`F1-Score:  ${(Number(f1) * 100).toFixed(2)}%\n`);6061const cm = confusionMatrix(y_test, y_pred);62console.log("Confusion Matrix:");63console.log(cm.toString());6465console.log("\n✓ Logistic regression complete!");

Console Output

$ npx tsx 08-logistic-regression/index.ts
=== Logistic Regression ===

Dataset: 150 samples, 4 features

Training set: 105 samples
Test set: 45 samples

Features scaled

Model trained!

Model Performance:
Accuracy:  100.00%
Precision: 100.00%
Recall:    100.00%
F1-Score:  100.00%

Confusion Matrix:
tensor([[13, 0]
       [0, 32]], dtype=float32)

✓ Logistic regression complete!
PreviousNext