Reader small image

You're reading from  Learning Jupyter

Product typeBook
Published inNov 2016
Reading LevelIntermediate
PublisherPackt
ISBN-139781785884870
Edition1st Edition
Languages
Tools
Right arrow
Author (1)
Dan Toomey
Dan Toomey
author image
Dan Toomey

Dan Toomey has been developing application software for over 20 years. He has worked in a variety of industries and companies, in roles from sole contributor to VP/CTO-level. For the last few years, he has been contracting for companies in the eastern Massachusetts area. Dan has been contracting under Dan Toomey Software Corp. Dan has also written R for Data Science, Jupyter for Data Sciences, and the Jupyter Cookbook, all with Packt.
Read more about Dan Toomey

Right arrow

Node.js decision-tree package


The decision-tree package is an example of a machine learning package. It is available at https://www.npmjs.com/package/decision-tree . The package is installed using the following command:

npm install decision-tree

We need a dataset to use for training/developing our decision tree. I am using the car MPG dataset on this page: https://alliance.seas.upenn.edu/~cis520/wiki/index.php?n=Lectures.DecisionTrees . It did not seem to be available directly, so I copied it into Excel and saved it as a local CSV.

The logic for machine learning is very similar:

  • Load our dataset

  • Split into a training set and a testing set

  • Use the training set to develop our model

  • Test the mode on the test set.

Tip

Typically, you might use two-thirds of your data for training and one-third for testing.

Using the decision-tree package and the car MPG dataset we would have a script similar to the following:

//Import the modules
var DecisionTree = require('decision-tree');
var fs = require("fs");...
lock icon
The rest of the page is locked
Previous PageNext Page
You have been reading a chapter from
Learning Jupyter
Published in: Nov 2016Publisher: PacktISBN-13: 9781785884870

Author (1)

author image
Dan Toomey

Dan Toomey has been developing application software for over 20 years. He has worked in a variety of industries and companies, in roles from sole contributor to VP/CTO-level. For the last few years, he has been contracting for companies in the eastern Massachusetts area. Dan has been contracting under Dan Toomey Software Corp. Dan has also written R for Data Science, Jupyter for Data Sciences, and the Jupyter Cookbook, all with Packt.
Read more about Dan Toomey