Recommenderlab question - want to predict and compare
up vote
0
down vote
favorite
I have product ratings for 30 users, lots of ratings per user
I have a realRatingMatrix, RLmatrix
RLmatrix
30 x 4592 rating matrix of class ‘realRatingMatrix’ with 19260 ratings.
I split into "train", "known", and "unknown" like so, using the first 200 ratings for each user as known:
eval_sets <- evaluationScheme(data = RLmatrix, method = "split", train = .667, given = 200, goodRating = 3.5, k=1)
getData(eval_sets,"train")
20 x 4592 rating matrix of class ‘realRatingMatrix’ with 13540 ratings.
getData(eval_sets,"known")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 2000 ratings.
getData(eval_sets,"unknown")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 3720 ratings.
I built a recommender on"train":
eval_recommender <- Recommender(data = getData(eval_sets, "train"), method = "IBCF", parameter= model_parameters)
I want to predict topNList, n=5, on the "known" set for the 10 test users, and see how many of these products were rated in the "unknown" set; I did this for the first test user:
rec <- predict(eval_recommender, RLmatrix["1724"], n=5);
BUT, Recommenderlab will make no predictions on the "unknown" set - all predictions have to be outside the set of users in the 4592 lines! And the same for all test users.
How do I get Recommenderlab to give me Top 5 lists based on the "train" set and the "known" ratings for a test user, which can include products in the "unknown" set for that test user?
recommenderlab
asked Nov 10 at 10:46
Phil Rennert
1
add a comment |
up vote
0
down vote
favorite
I have product ratings for 30 users, lots of ratings per user
I have a realRatingMatrix, RLmatrix
RLmatrix
30 x 4592 rating matrix of class ‘realRatingMatrix’ with 19260 ratings.
I split into "train", "known", and "unknown" like so, using the first 200 ratings for each user as known:
eval_sets <- evaluationScheme(data = RLmatrix, method = "split", train = .667, given = 200, goodRating = 3.5, k=1)
getData(eval_sets,"train")
20 x 4592 rating matrix of class ‘realRatingMatrix’ with 13540 ratings.
getData(eval_sets,"known")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 2000 ratings.
getData(eval_sets,"unknown")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 3720 ratings.
I built a recommender on"train":
eval_recommender <- Recommender(data = getData(eval_sets, "train"), method = "IBCF", parameter= model_parameters)
I want to predict topNList, n=5, on the "known" set for the 10 test users, and see how many of these products were rated in the "unknown" set; I did this for the first test user:
rec <- predict(eval_recommender, RLmatrix["1724"], n=5);
BUT, Recommenderlab will make no predictions on the "unknown" set - all predictions have to be outside the set of users in the 4592 lines! And the same for all test users.
How do I get Recommenderlab to give me Top 5 lists based on the "train" set and the "known" ratings for a test user, which can include products in the "unknown" set for that test user?
recommenderlab
asked Nov 10 at 10:46
Phil Rennert
1
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01
add a comment |
up vote
0
down vote
favorite
up vote
0
down vote
favorite
I have product ratings for 30 users, lots of ratings per user
I have a realRatingMatrix, RLmatrix
RLmatrix
30 x 4592 rating matrix of class ‘realRatingMatrix’ with 19260 ratings.
I split into "train", "known", and "unknown" like so, using the first 200 ratings for each user as known:
eval_sets <- evaluationScheme(data = RLmatrix, method = "split", train = .667, given = 200, goodRating = 3.5, k=1)
getData(eval_sets,"train")
20 x 4592 rating matrix of class ‘realRatingMatrix’ with 13540 ratings.
getData(eval_sets,"known")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 2000 ratings.
getData(eval_sets,"unknown")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 3720 ratings.
I built a recommender on"train":
eval_recommender <- Recommender(data = getData(eval_sets, "train"), method = "IBCF", parameter= model_parameters)
I want to predict topNList, n=5, on the "known" set for the 10 test users, and see how many of these products were rated in the "unknown" set; I did this for the first test user:
rec <- predict(eval_recommender, RLmatrix["1724"], n=5);
BUT, Recommenderlab will make no predictions on the "unknown" set - all predictions have to be outside the set of users in the 4592 lines! And the same for all test users.
How do I get Recommenderlab to give me Top 5 lists based on the "train" set and the "known" ratings for a test user, which can include products in the "unknown" set for that test user?
recommenderlab
asked Nov 10 at 10:46
Phil Rennert
1
I have product ratings for 30 users, lots of ratings per user
I have a realRatingMatrix, RLmatrix
RLmatrix
30 x 4592 rating matrix of class ‘realRatingMatrix’ with 19260 ratings.
I split into "train", "known", and "unknown" like so, using the first 200 ratings for each user as known:
eval_sets <- evaluationScheme(data = RLmatrix, method = "split", train = .667, given = 200, goodRating = 3.5, k=1)
getData(eval_sets,"train")
20 x 4592 rating matrix of class ‘realRatingMatrix’ with 13540 ratings.
getData(eval_sets,"known")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 2000 ratings.
getData(eval_sets,"unknown")
10 x 4592 rating matrix of class ‘realRatingMatrix’ with 3720 ratings.
I built a recommender on"train":
eval_recommender <- Recommender(data = getData(eval_sets, "train"), method = "IBCF", parameter= model_parameters)
I want to predict topNList, n=5, on the "known" set for the 10 test users, and see how many of these products were rated in the "unknown" set; I did this for the first test user:
rec <- predict(eval_recommender, RLmatrix["1724"], n=5);
BUT, Recommenderlab will make no predictions on the "unknown" set - all predictions have to be outside the set of users in the 4592 lines! And the same for all test users.
How do I get Recommenderlab to give me Top 5 lists based on the "train" set and the "known" ratings for a test user, which can include products in the "unknown" set for that test user?
recommenderlab
recommenderlab
asked Nov 10 at 10:46
Phil Rennert
1
asked Nov 10 at 10:46
Phil Rennert
1
asked Nov 10 at 10:46
Phil Rennert
1
asked Nov 10 at 10:46
Phil Rennert
1
asked Nov 10 at 10:46
Phil Rennert
1
1
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01
add a comment |
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01
add a comment |
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53238160%2frecommenderlab-question-want-to-predict-and-compare%23new-answer', 'question_page');
);
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Correction: I was wrong about predictions being outside the 4592 lines: they have to be within that set, that's all Recommenderlab has. So for the first test user, Recommenderlab happily recommends products from the "unknown" set of the second, third,...tenth test user. It just won't touch the "unknown" set of the target user! And if I try to conceal that set from Recommenderlab, I get a mismatch by length - not 4592 lines - and it errors out.
– Phil Rennert
Nov 10 at 11:01