datamagic-ml v1.0.0

DataMagic-ML

A lightweight JavaScript library for essential feature engineering tasks. Provides utilities for normalization, standardization, one-hot encoding, and missing value handling. Designed for simplicity and performance in both Node.js and browser environments.

Features

• Min-Max Scaling: Normalize data to a specific range.
• Standardization: Transform data to have zero mean and unit variance.
• One-Hot Encoding: Convert categorical data into numerical format.
• Missing Value Handling: Replace missing values with mean, median, or a constant.

Installation

You can install datamagic-ml via npm:

npm install datamagic-ml

Or using yarn:

yarn add datamagic-ml

Usage

Importing the Library:

const { MinMaxScaler, StandardScaler, OneHotEncoder, CleanMissings } = require('datamagic-ml');

Min-Max Scaling

const scaler = new MinMaxScaler();
const data = [1, 2, 3, 4, 5];
scaler.fit(data);
console.log(scaler.transform(data));

Standardization

const stdScaler = new StandardScaler();
const data = [1, 2, 3, 4, 5];
stdScaler.fit(data);
console.log(stdScaler.transform(data));

One-Hot Encoding

const encoder = new OneHotEncoder();
encoder.fit(['red', 'green', 'blue']);
console.log(encoder.transform(['green', 'red', 'yellow', 'blue'])); 

Handling Missing Values

const testArray = [1, null, 3, 4, NaN, 6];
console.log(CleanMissings(testArray, 'mean'));
console.log(CleanMissings(testArray, 'median'));
console.log(CleanMissings(testArray, 'constant', 0));

License

Licensed under the MIT License.