Q: How to test the recommand system result correct rate? about rrecsys

[ludovikcoba]

rrecsys method is used to train an algorithm on your data. Usually you want to train the train set and evaluate it with your "ground truth" data, i.e a test set.
I just pushed a new branch (rrecsys-split, use it in developer mode), that would work on evaluating a specific train set with another specific test set.
use as follows:

train <- defineData(train_set, intScale = F)

r <- rrecsys(train, "ub", neigh = 10, simFunct = "Pearson", coRatedThreshold = 2)
  
rec <- recommendHPR(r, train,topN = 3)

#test set as a data.frame 
e <- evalRec(rec, train, test, positiveThreshold = 3)
  

At the moment this branch works only with the UB.
I am updating the code to work on the algorithms, if you have a preference on a specific algorithm, just comment here so I update it first.

[evenbily]

It's very inspiring . And so grateful for your work .
It explains the matter.