Skip to content

AnuragAnalog/Try-every-ML-algorithm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Try every ML algorithm

We as a Data Scientist are very lazy in trying out every different algorithm on a given dataset.

This web interface provides you a convinent way of switching between algorithms are seeing there results. With this now, you can apply machine learning models without writing a Single Piece of code.

The web application is written in streamlit.

Link to the web app is here

How it works

  • Open the App
  • Upload the dataset*
  • Inspect your data, if you wish
  • Select the features
  • Select the label
  • Make a Train Test split
  • Select an appropriate algorithm
  • Start tweaking the hyperparameters

This app expects a preprocessed dataset with all the NaN, Null values handled properly, One Hot encoded, and scaled

Demo

demo

Required Modules

  • Pandas
  • Streamlit
  • Scikit-learn

Getting a copy of this repo

Clone the repository before running any commands

$ git clone https://github.com/AnuragAnalog/Try-every-ML-algorithm.git
$ cd Try-every-ML-algorithm

Installation

Run the below command to install all the dependencies in your local machine to run the py script.

$ sudo pip3 install -r requirements.txt

Running the app

$ streamlit run app.py

Algorithms

  • Regression

    • Linear Regression
    • K Nearest Neighbours
    • Decision Trees
    • Random Forest
    • Ada Boost
    • Gradient Boosting
  • Classification

    • Logistic Regression
    • K Nearest Neighbours
    • Decision Trees
    • Random Forest
    • Ada Boost
    • Gradient Boosting

Want to contribute? then fork, develop, and create a pull-request

Future Work

  • Add an option to show the code which implements the above selected algorithm with the corresponding hyperparameters.
  • Added code for One Hotencoding.
  • Add more algorithms.
  • Add some functionality for preprocessing data too.