Evaluation

Task 1. Author Profiling in Arabic Tweets

The performance of your author profiling approach will be evaluated by accuracy. For each subtask (age, gender, language variety), we will calculate individual accuracies. Systems will be ranked by the joint accuracy (when age, gender and language variety are properly identified together).

Task 2. Deception Detection in Arabic Texts

The performance of your deception detection approach will be evaluated using the macro-averaged measures (precision, recall and F1-score). Systems will be ranked by F1-score.