Huge News!Announcing our $40M Series B led by Abstract Ventures.Learn More
Socket
Sign inDemoInstall
Socket

softmax

Package Overview
Dependencies
Maintainers
1
Versions
14
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

softmax - npm Package Compare versions

Comparing version 0.5.0 to 2.0.0

lib/Algorithm.js

38

index.js

@@ -1,37 +0,1 @@

var _ = require('lodash');
var BPromise = require('bluebird');
var debug = require('debug')('softmax');
var async = BPromise.method;
function Algorithm(options) {
var opts = options || {};
if (!(this instanceof Algorithm)) {
return new Algorithm(opts);
}
debug('init', opts);
this.arms = _.isUndefined(opts.arms) ? 2 : parseInt(opts.arms, 10);
this.gamma = _.isUndefined(opts.gamma) ? 1e-7 : parseFloat(opts.gamma);
this.tau = _.isUndefined(opts.tau) ? null : parseFloat(opts.tau);
if (this.arms < 1) {
throw new TypeError('invalid arms: cannot be less than 1');
} else if (this.gamma < 0) {
throw new TypeError('invalid gamma: cannot be less than 0');
} else if (!_.isNull(this.tau) && this.tau < 0) {
throw new TypeError('invalid tau: cannot be less than 0');
}
this.counts = Array.apply(null, Array(this.arms)).map(Number.prototype.valueOf, 0);
this.values = Array.apply(null, Array(this.arms)).map(Number.prototype.valueOf, 0);
}
Algorithm.prototype.load = async(require('./lib/load'));
Algorithm.prototype.reward = async(require('./lib/reward'));
Algorithm.prototype.select = async(require('./lib/select'));
Algorithm.prototype.serialize = async(require('./lib/serialize'));
module.exports = Algorithm;
module.exports = require('./lib/Algorithm');
{
"name": "softmax",
"description": "A softmax multi-armed bandit algorithm",
"version": "0.5.0",
"version": "2.0.0",
"license": "ISC",

@@ -11,3 +11,3 @@ "main": "index.js",

"promises-aplus",
"banditlab"
"banditlab/2.0"
],

@@ -28,21 +28,23 @@ "author": {

"type": "git",
"url": "git://github.com/kurttheviking/softmax.git"
"url": "git://github.com/kurttheviking/softmax-js.git"
},
"bugs": {
"url": "https://github.com/kurttheviking/softmax/issues"
"url": "https://github.com/kurttheviking/softmax-js/issues"
},
"homepage": "https://github.com/kurttheviking/softmax#readme",
"homepage": "https://github.com/kurttheviking/softmax-js#readme",
"dependencies": {
"bluebird": "3.3.0",
"bluebird": "3.3.5",
"debug": "2.2.0",
"lodash": "4.3.0"
"lodash": "4.12.0"
},
"devDependencies": {
"chai": "3.5.0",
"eslint": "1.10.3",
"eslint-config-airbnb": "5.0.0",
"mocha-eslint": "1.0.0",
"istanbul": "0.4.2",
"eslint": "2.9.0",
"eslint-config-airbnb": "9.0.1",
"eslint-plugin-import": "1.8.0",
"mocha-eslint": "2.0.2",
"istanbul": "0.4.3",
"mocha": "2.4.5",
"sinon": "1.17.3"
"mockery": "1.7.0",
"sinon": "1.17.4"
},

@@ -52,3 +54,4 @@ "scripts": {

"test": "node node_modules/mocha/bin/mocha ./test --recursive"
}
},
"tonicExampleFilename": "./opt/tonic.js"
}
softmax
=======
[![Build Status](https://travis-ci.org/kurttheviking/softmax.svg)](https://travis-ci.org/kurttheviking/softmax)
[![Build Status](https://travis-ci.org/kurttheviking/softmax-js.svg)](https://travis-ci.org/kurttheviking/softmax-js)
**A softmax algorithm for multi-armed bandit problems**
**A softmax multi-armed bandit algorithm**
This implementation is based on [<em>Bandit Algorithms for Website Optimization</em>](http://shop.oreilly.com/product/0636920027393.do) and related empirical research in ["Algorithms for the multi-armed bandit problem"](https://d2w9gswcdc2jtf.cloudfront.net/research/Algorithms+for+the+multi-armed+bandit+problem.pdf).
This implementation is based on [<em>Bandit Algorithms for Website Optimization</em>](http://shop.oreilly.com/product/0636920027393.do) and related empirical research in ["Algorithms for the multi-armed bandit problem"](http://www.cs.mcgill.ca/~vkules/bandits.pdf).

@@ -13,3 +13,3 @@

This module conforms to the [BanditLab/1.0 specification](https://github.com/banditlab/spec-js/blob/master/README.md).
This module conforms to the [BanditLab/2.0 specification](https://github.com/banditlab/spec-js/releases).

@@ -27,3 +27,3 @@

1. Create an optimizer with 3 arms and default [annealing](https://en.wikipedia.org/wiki/Simulated_annealing):
1. Create an optimizer with `3` arms and default [annealing](https://en.wikipedia.org/wiki/Simulated_annealing):

@@ -42,3 +42,3 @@ ```js

algorithm.select().then(function (arm) {
...
// do something based on the chosen arm
});

@@ -50,5 +50,3 @@ ```

```js
algorithm.reward(armId, value).then(function (n) {
...
});
algorithm.reward(arm, value);
```

@@ -59,3 +57,3 @@

#### `Algorithm([config])`
#### `Algorithm(config)`

@@ -66,12 +64,14 @@ Create a new optimization algorithm.

- `config` (Object, Optional): algorithm instance parameters
- `config` (Object): algorithm instance parameters
The `config` object supports three parameters:
- `arms`: (Number:Integer, Optional), default=2, the number of arms over which the optimization will operate
- `gamma`: the annealing (cooling) factor &ndash; defaults to 1e-7 (0.0000001)
- `tau`: the temperature (scaling) factor &ndash; 0 to Infinity, higher leads to more exploration
- `arms`: (Number:Integer, Optional), defaults to `2`, the number of arms over which the optimization will operate
- `gamma`: annealing factor, defaults to `1e-7` (`0.0000001`)
- `tau`: fixed temperature, `0` to `Infinity`, higher leads to more exploration
By default, `gamma` of 1e-7 will cause the algorithm to explore less as more information is received. In this case, the underlying "temperature" is changing. If this behavior is not desired, set `tau` to instead employ an algorithm with a fixed temperature. If `tau` is provided then `gamma` is ignored.
By default, `gamma` is set to `1e-7` which causes the algorithm to reduce exploration as more information is received. That is, the "temperature cools" slightly with each iteration. In contrast, `tau` represents a "constant temperature" wherein the influence of random search is fixed across all iterations. If `tau` is provided then `gamma` is ignored.
Alternatively, the `state` object returned from [`Algorithm#serialize`](https://github.com/kurttheviking/softmax-js-js#algorithmserialize) can be passed as `config`.
**Returns**

@@ -84,6 +84,7 @@

```js
> var Algorithm = require('softmax');
> var algorithm = new Algorithm();
> assert.equal(algorithm.arms, 3);
> assert.equal(algorithm.gamma, 0.0000001);
var Algorithm = require('softmax');
var algorithm = new Algorithm();
assert.equal(algorithm.arms, 3);
assert.equal(algorithm.gamma, 0.0000001);
```

@@ -94,6 +95,7 @@

```js
> var Algorithm = require('softmax');
> var algorithm = new Algorithm({arms: 4, tau: 0.000005});
> assert.equal(algorithm.arms, 4);
> assert.equal(algorithm.tau, 0.000005);
var Algorithm = require('softmax');
var algorithm = new Algorithm({arms: 4, tau: 0.000005});
assert.equal(algorithm.arms, 4);
assert.equal(algorithm.tau, 0.000005);
```

@@ -116,6 +118,9 @@

```js
> var Algorithm = require('softmax');
> var algorithm = new Algorithm();
> algorithm.select().then(function (arm) { console.log(arm); });
var Algorithm = require('softmax');
var algorithm = new Algorithm();
algorithm.select().then(function (arm) { console.log(arm); });
```
```js
0

@@ -135,3 +140,3 @@ ```

A promise that resolves to a Number representing the count of observed rounds.
A promise that resolves to an updated instance of the algorithm.

@@ -141,9 +146,17 @@ **Example**

```js
> var Algorithm = require('softmax');
> var algorithm = new Algorithm();
> algorithm.reward(0, 1).then(function (n) { console.log(n); });
var Algorithm = require('softmax');
var algorithm = new Algorithm();
1
algorithm.reward(0, 1).then(function (algorithmUpdated) { console.log(algorithmUpdated) });
```
```js
<Algorithm>{
arms: 2,
gamma: 0.0000001,
counts: [1, 0],
values: [1, 0]
}
```
#### `Algorithm#serialize()`

@@ -164,6 +177,9 @@

```js
> var Algorithm = require('softmax');
> var algorithm = new Algorithm();
> algorithm.serialize().then(function (state) { console.log(state); });
var Algorithm = require('softmax');
var algorithm = new Algorithm();
algorithm.serialize().then(function (state) { console.log(state); });
```
```js
{

@@ -177,26 +193,3 @@ arms: 2,

#### `Algorithm#load(state)`
Restore an instance of an algorithm to a previously serialized state. This method overrides any options parameters passed at instantiation.
**Arguments**
- `state` (Object): a serialized algorithm state (provided from `algorithm.serialize()`)
**Returns**
A promise that resolves to a Number representing the count of observed rounds.
**Example**
```js
> var state = {arms: 2, gamma: 0.0000001, counts: [1, 2], values: [1, 0.5]};
> var Algorithm = require('softmax');
> var algorithm = new Algorithm();
> algorithm.load(state).then(function (n) { console.log(n); });
3
```
## Tests

@@ -221,3 +214,3 @@

PRs are welcome! For bugs, please include a failing test which passes when your PR is applied. [Travis CI](https://travis-ci.org/kurttheviking/softmax) provides on-demand testing for commits and pull requests.
PRs are welcome! For bugs, please include a failing test which passes when your PR is applied. [Travis CI](https://travis-ci.org/kurttheviking/softmax-js) provides on-demand testing for commits and pull requests.

@@ -224,0 +217,0 @@

Sorry, the diff of this file is not supported yet

Sorry, the diff of this file is not supported yet

SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

Packages

npm

Stay in touch

Get open source security insights delivered straight into your inbox.


  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc