PredictionIO Template: House Prices

Overview

This repository is PredictionIO Template for House Prices: Advanced Regression Techniques on Kaggle.

Dataset

Download train.csv and test.csv from Data and put them to data directory.

Requirement

• Docker

Getting Started

Run Jupyter

PIO_MODE=jupyter docker-compose up --abort-on-container-exit

Open eda.ipynb at Jupyter Notebook page(http://localhost:8888/).

Run Train Steps on Docker

train mode executes create-app, import-data, build and train process.

PIO_MODE=train docker-compose up --abort-on-container-exit

Run Deploy Step on Docker

deploy mode starts Rest API on Docker container.

PIO_MODE=deploy docker-compose up --abort-on-container-exit

and then check a prediction:

curl -s -H "Content-Type: application/json" -d '{"attr0":5.1,"attr1":3.5,"attr2":1.4,"attr3":0.2}' http://localhost:8000/queries.json