Research Article

Improving Feature Representation Based on a Neural Network for Author Profiling in Social Media Texts

Table 14

Obtained results (accuracy, %) for gender classification using SVM classifier on the PAN author profiling 2015 Dutch training corpus under 10-fold cross-validation with vectors extracted only from the training data.

Feature set Gender
SVM-NPSVM-WP

D2V (1-gram) 72.5075.00
D2V (1 + 2-grams) 65.00 70.00
D2V (1 + 2 + 3-grams) 65.00 72.50