Overview / Usage

What is RapidML?

Well, RapidML is your Smart Machine Learning assistant that not only automates the creation of machine learning models but also enables you to easily deploy the models to the cloud.

RapidML is perfect for Python programmers at all levels, ranging from beginners who want to get into Data Science and Machine Learning to intermediate and advanced programmers who want to bring Machine Learning to consumer and industry usage applications.

Apart from making predictions in Python, RapidML models can be exported as Web APIs to develop Machine Learning applications in a wide variety of platforms, such as Javascript, Android, iOS…. and almost everything else which can make and receive web requests!

Methodology / Approach

RapidML does a lot of the heavy lifting in machine learning for you.

1. Unlike other machine learning libraries which require X,y arrays for training data and target, RapidML can take data frame inputs which results in fewer hassles for the programmer.

2. Furthermore, RapidML auto-encodes your data because many ML algorithms require numerical data. (So textual categorical data will like genders: “male”/“female” will fail while numerical values: 0/1 will work).

3. Where RapidML truly shines is that it automatically selects the best machine learning algorithm and performs hyperparameter optimization based on your data.

4. Furthermore, it saves the machine learning model, a Flask API and a bunch of helper files to a directory of your choice. Now all you have to do is to upload this directory to a WSGI server and you have a functioning machine learning API.

Technologies Used

1. Python
2. Flask
3. Scikit-learn
4. TPOT
5. Dill
6. Numpy, pandas etc. for data wrangling

Repository

https://github.com/ritabratamaiti/RapidML

Other links

• Documentation (Mirror Source)
• Documentation