CODEWITHSUNDEEP
machine-learningmahoutrecommendation-enginecollaborative-filtering
Ashika Umanga Umagiliya
Ashika Umanga Umagiliya

Reputation: 9178

Location aware recommendation system (Collaborative Filtering)?

I am aware that its possible to build a recommendation system with Mahout, But is it possible to make it location aware?

For example, first it filter out the nodes within a certain radius (using geospatial data) and then apply recommender on those data (Mahout SVDRecommender)

Upvotes: 0

Views: 150

Answers (1)

Othman Darraz
Othman Darraz

Reputation: 21

When you make a recommandation, you can provide a IDRescorer :

public List<RecommendedItem> recommend(long userID, int howMany, IDRescorer rescorer) throws TasteException

An IDRescorer is an interface that you can implement in order to filter out or rescore depending on criterias you define yourself. Please refer to Mahout documentation for more details on the interface :

https://archive.cloudera.com/cdh4/cdh/4/mahout-0.7-cdh4.5.0/mahout-core/org/apache/mahout/cf/taste/recommender/IDRescorer.html

In your case, you could filter out nodes that will not be in that radius, or rescore by assigning higher scores to those in a certain radius.

Upvotes: 2

Related Questions