katungi / KNN-TensorFlow-JS

K-Nearest-Neighbor Algorithm implemented in TensorFlow JS ML Library

KNN-TensorFlow-JS

K-Nearest-Neighbor Algorithm implemented in TensorFlow JS ML Library. Coded along with Stephen Grider in Machine Learning with JavaScript Udemy Course (https://www.udemy.com/machine-learning-with-javascript/).

**Overview of KNN Algorithm

*Housing Data Set (CSV file) is included

*Features are latitude, longitude, lot (in sqft), living (in sqft)

*Label and Prediction Point is Housing Prices (in Thousands of $s)

Steps:

**Download NPM Dependencies:

1. Please clone or download this repo on to your desktop
2. Open up a terminal window and CD into knn-tf directory
3. Run 'npm install --force' in the terminal window pointing to the knn-tf directory { to download npm modules and dependencies (Package.json files are included in repo)}

**To Run KNN Algorithm in your CPU:

4. Run 'node index.js' in the terminal window pointing to the knn-tf directory to execute KNN algorithm and run Tensorflow JS in Node
5. You should see KNN Label Results (Housing Prices in Thousands of $s) and KNN Error Percentages (to show discrepancies between observed predictions by KNN algorithm and actual testing results)