Distance Measures, KNN, Recommendation

Need to compute distance measures: Manhattan, Euclidean, Pearson and Cosine and do the KNN algorithm using all the distances to see which person is more likely to rate the movies to another person. And also see which measure’s the best among all.
Then create a recommendation system to suggest movies (out of those 8 movies in the dataset Movies.xlsx (9.0 KB)) to all 8 people.

And because I need to keep the missings to check which of them distances are better, I don’t want to remove or replace the missings, so any hint or suggestion?!


Hi @HamidR3zZ and welcome to the forum.

If you want to calculate the distances you mention in KNIME you can use the Distance Matrix Calculate node, which can then in turn be used as input to the K Nearest Neighbor (Distance Function) node.

If you’re interested in other recommendation approaches, you might also try this example workflow, which uses the Association Rule Learner (Borgelt) node: https://kni.me/w/q0znOkLe5WCf47b7

Thanks a million, I’ll give it a try and share the experience as well
all the best

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.