The Root Mean Squared Error ("RMSE") is used to measure the accuracy:
\[ \textrm{RMSE} = \sqrt{\frac{\sum_{i=1}^n (p_i - a_i)^2}{n} } \]
Where:
\\( n \\) is 24 times the total number of tweets
\\( p_i \\) is the predicted confidence rating for a given label
\\( a_i \\) is the actual confidence rating for a given label
Submission format
For each tweet in the test set, predict the confidence score for each of the 24 possible labels.  Submission files should be in the following format and must have a header.
id,s1,s2,s3,s4,...,k12,k13,k14,k15
4,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
5,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
7,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
etc.