• Introduction

      To see how PredictionIO works, you can follow this quick tutorial to:

      Start the PredictionIO server

      Build, train, and deploy a predictive model with pre-imported data.

      Query prediction results.

      Start

    • Start

      Start PredictionIO and it's dependencies with:

      $ pio-start-all

      Check Status

      At any time you can run:

      $ pio status

      Which checks the status of PredictionIO and it's dependencies.

    • Build

      All engine commands need to be run from within the engine directory:

      $ cd ~/PredictionIO/MyEngine

      Build the engine with:

      $ pio build

    • Train and Deploy

      Train the engine

      Sample data is already imported, you can train the model the data by running:

      $ pio train

      Deploy the trained engine

      Deploy the trained engine as a background process:

      $ pio deploy &

    • Querying Results

      You can query results using cURL with after you see "Ready to serve":

      $ curl -H "Content-Type: application/json" -d '{ "items": ["296"], "num": 5 }' http://localhost:8000/queries.json

      This will return 5 recommended movies for movie 296 (Pulp Fiction).

    • Next Steps

      Install PredictionIO on your own computer!

      Support

      Get support on our Google Group or on Twitter @PredictionIO

    ‹ PrevNext ›