accuracy-meter

accuracy-meter

Measure accuracy, precision and recall of a machine learning classifier.

Usage

var meter = new AccuracyMeter();

dataset.forEach(function(){

	meter.add(['positive'],['negative']); // meter.add(predicted, golden);

});

meter.get();

API

add(predicted, golden)

It adds a pair <predicted, golden>, representing the set of prdicted labels and the set of golden labels. Golden and predicted labels are both arrays.

get()

It returns an object containing:

• precision
• recall
• fscore

Why

• node.js is not a standard language for NLP tasks, but...
• machine learning is more and more in the cloud (check Watson NLP Classifier)
• machine learning is becoming just a set of Http end points to invoke
• node.js is perfect as middleware and client among Http end points
• ...accuracy-meter might help you to measure the accuracy of your algorithms in the cloud