Page 1 of 1

Recommendation algorithms

Posted: Wed Apr 06, 2022 6:38 am
by Lin
Hi i want to know if you intend to add some implementations of recommandation algorithms to your software?

Re: Recommendation algorithms

Posted: Wed Apr 06, 2022 1:20 pm
by admin
Hi there,

welcome to the forum.

Currently there is no plan for the implementation of recommendation algorithms for recommender systems.

However, there are some algorithms offered in SPMF that could be used for recommendation or as a component in a recommender system.

For example, some researchers have used the CPT and CPT+ model of SPMF to perform sequence prediction or to predict the next locations of cars, and other similar tasks.

Hope this helps.

Philippe

Re: Recommendation algorithms

Posted: Wed Apr 06, 2022 11:26 pm
by gmc
I saw also that SPMF has algorithms for classification using association rules. That may not five the best classification accuracy nowadays But could be used

Re: Recommendation algorithms

Posted: Wed Apr 13, 2022 3:03 am
by admin
Yes, using association rules for classification may not give the best accuracy in many cases but can be competitive to some traditional classification algorithms like decision trees in some cases. But an advantage of rule based or tree based classifiers is that they sometimes can be interpretable while some other models are difficult to interpret. So I would say whether to use such models or not depends on what you want to do, how much data you have, how the data is, and whether the rule format is interesting for you.