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

Table 6

Obtained results (accuracy, %) for age and gender classification on the PAN author profiling 2015 Spanish training corpus under 10-fold cross-validation.

Feature set	Age
	LR-NP	LR-WP	SVM-NP	SVM-WP
D2V (1-gram)	59.00	62.00	62.00	60.00
D2V (1 + 2-grams)	59.00	65.00	65.00	69.00
D2V (1 + 2 + 3-grams)	62.00	66.00	64.00	66.00
Character 3-grams	66.00	66.00	64.00	67.00
Bag-of-Words	65.00	62.00	62.00	60.00
Feature set	Gender
	LR-NP	LR-WP	SVM-NP	SVM-WP
D2V (1-gram)	65.00	63.00	63.00	66.00
D2V (1 + 2-grams)	68.00	66.00	66.00	61.00
D2V (1 + 2 + 3-grams)	71.00	67.00	71.00	66.00
Character 3-grams	73.00	73.00	75.00	74.00
Bag-of-; Words	72.00	71.00	73.00	72.00