@ai-on-browser/data-analysis-models/lib/model/sop.js

Data analysis model package without any dependencies

import Matrix from '../util/matrix.js'

/**
 * Second order perceptron
 */
export default class SecondOrderPerceptron {
	// http://www.datascienceassn.org/sites/default/files/Second-order%20Perception%20Algorithm.pdf
	// A SECOND-ORDER PERCEPTRON ALGORITHM. (2005)
	/**
	 * @param {number} [a] Tuning parameter
	 */
	constructor(a = 1) {
		this._v = null
		this._s = null
		this._a = a
	}

	_w() {
		const s = Matrix.eye(this._d, this._d, this._a)
		s.add(this._s.dot(this._s.t))
		return s.solve(this._v)
	}

	/**
	 * Initialize this model.
	 * @param {Array<Array<number>>} train_x Training data
	 * @param {Array<1 | -1>} train_y Target values
	 */
	init(train_x, train_y) {
		this._x = Matrix.fromArray(train_x)
		this._shift = this._x.mean(0)
		this._x.sub(this._shift)
		this._y = train_y

		this._d = this._x.cols
		this._v = Matrix.zeros(this._d, 1)
		this._s = Matrix.zeros(this._d, 0)
	}

	/**
	 * Update model parameters with one data.
	 * @param {Matrix} x Training data
	 * @param {1 | -1} y Target value
	 */
	update(x, y) {
		this._s.concat(x, 1)
		const w = this._w()
		const m = w.tDot(x).toScaler()
		if (m * y > 0) return

		this._v.add(Matrix.mult(x, y))
	}

	/**
	 * Fit model parameters.
	 */
	fit() {
		for (let i = 0; i < this._x.rows; i++) {
			this.update(this._x.row(i).t, this._y[i])
		}
	}

	/**
	 * Returns predicted datas.
	 * @param {Array<Array<number>>} data Sample data
	 * @returns {(1 | -1)[]} Predicted values
	 */
	predict(data) {
		const x = Matrix.fromArray(data)
		x.sub(this._shift)
		const r = x.dot(this._w())
		return r.value.map(v => (v <= 0 ? -1 : 1))
	}
}