irecsys
commented
5 years ago The option of "NA" is only applied when you have missing values in the data. CARSKit will automatically add this context condition in data transformations. A rating with all "NA" conditions is referred to a rating without considering contexts.
The input format could be binary, compact or loose format. Please refer to the CARSKit guide. Your data format looks good to me.
Hi, 1)Please, I wanted to know, is it MANDATORY to have some NA values in the context fields if I am planning to use similarity CAMF? 2) If the answer in 1 is yes, how should I assign them in my dataset, just randomly change some context condition to be NA? 3)Kindly can you confirm if the format below is correct for the input file as I get 1 all the time in REC10 when I choose CAMF models?
userid,itemid,rating,p1,p2,p3,p4 1,1,3,NA,NA,NA,NA 1,1,3,NA,NA,NA,NA 1,1,3,NA,NA,NA,NA 1,1,3,NA,NA,NA,NA 1,1,4,X-Large,X-Small,Nominal,Small 1,1,3,NA,NA,NA,NA 1,1,2,X-Large,X-Large,Numeric,X-Large 1,1,3,NA,NA,NA,NA 1,1,3,NA,NA,NA,NA 1,1,3,NA,NA,NA,NA 1,1,2,NA,NA,NA,NA 1,1,2,X-Large,X-Large,Numeric,X-Large 1,1,2,X-Large,X-Large,Numeric,X-Large 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,4,X-Large,X-Small,Numeric,X-Small 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,5,NA,NA,NA,NA 1,1,4,Large,X-Small,Numeric,X-Small 1,1,4,X-Large,X-Small,Numeric,X-Small 1,1,4,X-Large,X-Small,Numeric,X-Small 1,1,4,Large,X-Small,Numeric,X-Small 1,1,5,NA,NA,NA,NA Thank you.