How do I make a recommender system for a movie stream service

alisajjadhasan

Hello ! I am new to the community and I need some help in making a basic movie recommendation system as a school project. Can someone please help me with the basics of this as i am new to this software.

 

Thanks!

Thomas_Ott

Download the Recommender Extension here: https://marketplace.rapidminer.com/UpdateServer/faces/product_details.xhtml?productId=rmx_irbrecommender

 

Then visit http://elico.rapid-i.com/recommender-extension.html and find they're sample workflows.

 

 

alisajjadhasan

Thank you Thomas, but out instructor didnot provide information about this extension that you have mentioned. Is it possible to make a recommendation workflow without the extension ? if so what would be the difference in it ?

Thomas_Ott

The Recommender Extension has a few more algorithms that are not native to the RapidMiner platform. I'm not sure if you can build a recommendation engine without the extension. I've never built one.

alisajjadhasan

thank you very much for the replies and i am sorry to bug you like this but this means a lot to me. let me clarify the task a bit more and if you can please enlighten me with some info on how to do it. this is the task that has been given to us,

 

Develop a recommender system for a movie streaming service.
Using Rapid Miner you need to create and test a recommendation system that is as accurate as possible

 

we have been provided with 4 data sets (.CSV Files) which are LINKS, MOVIES, TAGS and RATINGS. they all have thounsands of entries and i am a little confused on how to work with this.

 

suggestive readings are,

 

Constructing recommender systems workflow templates in RapidMiner

 

A Decision Tree Based Recommender System

 

Extending RapidMiner with recommender systems algorithms

 

 

your help will be greatly appreciated sir !

Thomas_Ott

"Extending RapidMiner with recommender systems algorithms" and "Constructing recommender systems workflow templates in RapidMiner"

 

I read that to mean the Recommender Extension and sample workflows, as I proposed above. 

Telcontar120

Also, the recommender system extension is free, so there really isn't a reason not to try it.  It sounds like you are going to have to do some preliminary data ETL to combine the information from your 4 separate files into a single dataset and then use that dataset with the recommender operators to create a recommender system, along the lines described in the resources that Tom has suggested.