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

Table 9

Obtained results (accuracy, %) for age and gender classification on the PAN author profiling 2016 English training corpus under 10-fold cross-validation.

Feature set	Age
	LR-NP	LR-WP	SVM-NP	SVM-WP
D2V (1-gram)	44.71	44.84	41.78	41.65
D2V (1 + 2-grams)	43.53	44.37	42.82	42.49
D2V (1 + 2 + 3-grams)	41.41	46.71	40.71	44.13
Character 3-grams	39.53	41.78	37.65	42.96
Bag-of-Words	42.82	39.44	40.94	39.91
Feature set	Gender
	LR-NP	LR-WP	SVM-NP	SVM-WP
D2V (1-gram)	73.18	75.59	72.71	70.66
D2V (1 + 2-grams)	73.41	78.64	71.53	74.88
D2V (1 + 2 + 3-grams)	71.53	77.46	69.41	76.76
Character 3-grams	68.47	73.24	69.65	71.83
Bag-of-Words	69.18	72.77	67.76	69.01