https://github.com/mudafar/alz-predictor

A JavaScript sequence prediction library, for building proactive and smart user interfaces.
https://github.com/mudafar/alz-predictor

activelezi adaptive ai algorithm client-side javascript-library lz77 prediction predictive-prefetching sequence spa ui

Last synced: about 7 hours ago
JSON representation

A JavaScript sequence prediction library, for building proactive and smart user interfaces.

• Host: GitHub
• URL: https://github.com/mudafar/alz-predictor
• Owner: mudafar
• License: mit
• Created: 2021-03-22T22:50:46.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2021-04-03T17:08:49.000Z (over 3 years ago)
• Last Synced: 2024-04-27T01:43:05.557Z (7 months ago)
• Topics: activelezi, adaptive, ai, algorithm, client-side, javascript-library, lz77, prediction, predictive-prefetching, sequence, spa, ui
• Language: JavaScript
• Homepage:
• Size: 173 KB
• Stars: 9
• Watchers: 1
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# ALZ predictor · [![GitHub license](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/mudafar/alz-predictor/blob/master/LICENSE) [![npm version](https://img.shields.io/npm/v/alz-predictor.svg?style=flat)](https://www.npmjs.com/package/alz-predictor)  ![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)

ALZ predictor is a JavaScript library for symbols sequence prediction, based on **Gopalratnam and Cook Active LeZi algorithm**.

* **AI:** ALZ predictor makes it possible to create proactive and smart UIs. Improve your application UX by adding a touch of artificial intelligence. Guess and prefetch required data to enhance performance. 

* **Simple:** Model each user action as a char, then train ALZ by simply adding them as they come (one by one). Predict next user action at any time. 

* **Lightweight:** No dependencies.   

## FAQ

- Why **JavaScript** and **Frontend AI** at all? 

   - No server-client latency

   - Offline

   - Privacy

   - Cost

   - [TensorFlow.js](https://www.tensorflow.org/js/) and [ONNX.JS](https://microsoft.github.io/onnxjs-demo/#/)

- How to **save learning** data?

   

   Serialize and save to browser **localStorage**.

- What is the difference with **Guess.js**?

   

    **Google Analytics** is **not** required for training.

## Installation

```bash

# Yarn

yarn add alz-predictor

# NPM

npm install --save alz-predictor

```

## Usage

```js

import Predictor from 'alz-predictor'

const predictor = new Predictor()

predictor.add('a')

predictor.add('b')

predictor.add('a')

const predictions = predictor.predict()

```

## Examples

- Next-page data predective prefetching:

    - Online demo: https://mudafar.github.io/alz-predictor-prefetch-example/

    - [Source code](https://github.com/mudafar/alz-predictor-prefetch-example)

## Documentation

### Constructor

```js

const predictor = new Predictor()

```

### Methods

- add: add `char` to the sequence.

- predict: get predictions object. 

- loadJSON: recover state from JSON. 

### Serialization

```js

const predictor = new Predictor()

predictor.add('a')

predictor.add('b')

const json = JSON.stringify(predictor)

```

## License

ALZ predictor is [MIT licensed](./LICENSE).