In statistical classification, the accuracy of a classifier is the ratio between the number of correct predictions and the number of predictions.
$$ \text{accuracy} = \frac{N_\text{correctly predicted}}{N}. $$
See also Recall (classification), Precision (classification) and F-score.