The SitePoint PHP blog has posted the next part in their series about using Predictive.IO and Lumen to create a simple movie recommendation application (part one is here). In this second part of the series they build on the environment created in the previous article and start developing the actual application.
In the intro, we covered the basics of PredictionIO and installed its dependencies. In this part, we’re going to build the movie recommendation app.
The tutorial starts with a brief configuration section to ensure you have your API keys configured correctly. Then it gets into the code:
- Pulling the data from the Movie DB API
- Creating the endpoint to perform the endpoint
- Picking random movies to show the user and recording their reactions (like/dislike)
- Creating the endpoint to recommend movies
Finally they share the configuration to set up the application deployment and train it with some example content you provide through some basic interactions. Finally they help you set up a cron job to train and redeploy the application every five minutes with the latest interaction information.