Machine Learning and AI experiments using Node.js as server, client-side JavaScript operations and and MongoDB as data source.
../data/ folder contains the example database, called 'dataset', with a training set in every collection.
All the important stuff is client side and is in the ../public/javascripts/ folder, Node.js is only used as a server and to retreive data from MongoDB.
You really don't. This is meant to be just an experiment or simply practice, and there are a lot of libraries for machine learning out there that are way better then this one.
- a simple C4.5
- a simple K-Means
- Cross Validation
- Simulated Annealing
- Genetic Algorithm
##Installation To use the main functions you really just need a browser and the stuff in the ../public/javascripts/ folder. Just give it a data set in JSON format (se demo.js or tests.html for some examples).
For server and database functions you must first install Node.js and MongoDB. Then simply open a terminal and run this command:
cd /your/path/to/Machine-Learning-JS/
npm install
It has to be done only the first time in order to install the required modules.
Then you should run one of those commands in a terminal (depending on your OS):
#(Unix)
/your/path/to/mongod --dbpath /your/path/to/Machine-Learning-JS/data
::(Windows)
"Drive:\your\path\to\mongod.exe" --dbpath "Drive:\your\path\to\Machine-Learning-JS\data"
And this one in another terminal to start the server (of course current directory has to be the project folder again):
cd /your/path/to/Machine-Learning-JS/
npm start
Finally go to localhost:3000 in your browser and wait for the data to be loaded (it may take a while depending on your hardware).
Base for pseudo-code and many ideas are directly derived from my course book, "Artificial Intelligence: A Modern Approach (Third edition) by Stuart Russell and Peter Norvig". That book is simply amazing.
The pseudo-code I used for my K-Means implementation is taken from Stanford.edu