Stay Hungry,Stay Foolish!

a simple machine learning system demo, for ML study.

Machine Learning System

introduction

This project is a full stack Django/React/Redux app that uses token based authentication with Knox.

Then I add Machine Learning features for demostrate the full workflow of the data mining, including the four stage corresponding to four pages:

  1. data management
  2. data explore
  3. model train
  4. prediction

The data set is the classic iris data, which is only for demo, and this project is from my interest. so you can reference, but the quality is not assured.

 

features

  • authentication functions

login from login page register your account logout from inner page

  • data management

input iris items edit iris items delete iris items

  • data explore

inspect attribute distribution through histogram inspect sepal distribution through scatter graph inspect petal distribution through scatter graph

  • model train

input cluster number train a cluster model using sklearn-kmeans library inspect cluster result through sepal and petal scatter

  • prediction

input iris sepal and petal attributes predict iris cluster

 

technology stack

categorynamecomment
frontend reactjs frontend framework
frontend redux state management
frontend react-C3JS D3 based graph tool
frontend react-bootstrap style component library
frontend data-ui react data visualization tool
backend django backend framework
backend django-rest-knox authentication library
backend djangorestframework restful framework
backend sklearn machine learning tool


Quick Start

# Install dependencies
cd ./frontend
npm install

# Build for production
npm run build


# Install dependencies
cd ../backend
pipenv install

# Serve API on localhost:8000
pipenv run python manage.py runserver

 

snapshot

 

login page

avatar

 

model train page

avatar

 

prediction page

avatar

 

posted @ 2019-05-01 21:58  lightsong  阅读(228)  评论(0编辑  收藏  举报
Life Is Short, We Need Ship To Travel