The evaluation metric for this competition is prediction accuracy (percentage of correctly classified missing values).
Submission Format
For every missing value in the dataset
, submission file should contain two columns: Id (integer) and Prediction (0/1). The Id is constructed as follows:
\[ Id = i+(j-1)*N_{rows}, \]
where i is the row index (ranging between 1 and N
rows
) and j is the column index. Matlab, for example, uses the same indexing. The resulting indexing should look like this if the matrix has 3 rows and 4 column
\[ \left( \begin{array}{cccc}
1 & 4 & 7 & 10 \\
2 & 5 & 8 & 11\\
3 & 6 & 9 & 12 \end{array} \right)\]
Note that eventhough the provided user-artist matrix contains values >1, you are expected to
submit only 0/1 predictions
! Value 1 means that you believe that the user has listened to at least one song from the artist.
The file should contain a header and have the following format:
Id,Prediction
81,1
916,0
1138,1
1203,0
etc.