https://github.com/math-utils/wilson-score

Because averages suck
https://github.com/math-utils/wilson-score

Last synced: 10 days ago
JSON representation

Because averages suck

• Host: GitHub
• URL: https://github.com/math-utils/wilson-score
• Owner: math-utils
• License: mit
• Created: 2013-10-11T00:56:08.000Z (about 11 years ago)
• Default Branch: master
• Last Pushed: 2020-07-23T12:28:41.000Z (over 4 years ago)
• Last Synced: 2024-10-28T09:53:30.534Z (17 days ago)
• Language: JavaScript
• Size: 5.86 KB
• Stars: 41
• Watchers: 1
• Forks: 4
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Wilson Score [![Build Status](https://travis-ci.org/math-utils/wilson-score.png)](https://travis-ci.org/math-utils/wilson-score)

Calculates the lower bound of the [binomial proportion confidence interval](http://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval) as calculated by the [Wilson score interval](http://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval#Wilson_score_interval).

If you're not retarded, you would use this score instead of averages when sorting and ranking (doesn't matter what the user sees).

This is what [reddit uses](http://amix.dk/blog/post/19588) for "best".

## API

```js

var wilson = require('wilson-score')

```

### wilson(up, total [, z-score])

- `up` - the total number of "yes"s

- `total` - the total number of votes

- `z-score` - the z-score of the interval.

  By default, it's ~2.3 which corresponds to a 99% confidence interval.

  Use `1.644853` for 95%.

Instead of calculating the average as `up / total`,

calculate the wilson score via `wilson(up, total)`.

The wilson score will always be lower than the average.