LDA

Linear Discriminant Analysis (LDA) is a technique used to find a linear combination of features that characterizes or separates two or more classes of objects or events.

How It Works

Linear Discriminant Analysis (LDA) finds a linear combination of features that maximizes the ratio of between-class variance to within-class variance in the dataset.

Why or When to Use

Use LDA for supervised dimensionality reduction where class labels are known, aiming to separate distinct classes as much as possible, often as a preprocessing step for classification.

Example

How-to (Code)

import * as druid from "@saehrimnir/druidjs";

const data = [
  /* ... multi-dimensional data ... */
];
const classLabels = [
  /* ... labels ... */
];

// 1. Initialize the algorithm
const lda = new druid.LDA(data, { labels: classLabels });

// 2. Compute the projection
const projection = lda.transform();