Multilingual Hope Speech Detection (HOPE)

The HOPE task (webpage) presented at IberLEF 2023 focused on the identification of positive speech in Spanish and English.

The following code can generate an instance of the system used in the competition.

>>> from EvoMSA.competitions import Comp2023
>>> D = # Training set
>>> comp2023 = Comp2023(lang='es')
>>> ins = comp2023.stack_bow_keywords_emojis_voc_selection(D)

Performance in Cross-validation (Spanish)

Configuration	Performance	p-value
Comp2023.stack_bow_keywords_emojis_voc_selection	0.8224	1.0000
Comp2023.stack_bow_keywords_emojis	0.8217	0.3580
Comp2023.stack_3_bows_tailored_keywords	0.8192	0.3680
Comp2023.stack_2_bow_all_keywords	0.8192	0.3040
Comp2023.stack_3_bow_tailored_all_keywords	0.8192	0.3680
Comp2023.stack_2_bow_tailored_all_keywords	0.8159	0.1740
Comp2023.stack_2_bow_keywords	0.8152	0.1500
Comp2023.stack_2_bow_tailored_keywords	0.8119	0.1020
Comp2023.stack_3_bows	0.7998	0.0460
Comp2023.bow	0.7966	0.0260
Comp2023.stack_bows	0.7966	0.0260
Comp2023.bow_voc_selection	0.7902	0.0040
Comp2023.bow_training_set	0.7807	0.0000

The following code can generate an instance of the system used in the competition.

>>> from EvoMSA.competitions import Comp2023
>>> D = # Training set
>>> comp2023 = Comp2023(lang='en')
>>> ins = comp2023.stack_bow_keywords_emojis(D)

Performance in Cross-validation (English)

Configuration	Performance	p-value
Comp2023.stack_bow_keywords_emojis	0.7807	1.0000
Comp2023.stack_bow_keywords_emojis_voc_selection	0.7801	0.4600
Comp2023.stack_bows	0.7753	0.2860
Comp2023.stack_2_bow_tailored_keywords	0.7691	0.0300
Comp2023.stack_2_bow_tailored_all_keywords	0.7690	0.0260
Comp2023.stack_2_bow_keywords	0.7680	0.0220
Comp2023.stack_2_bow_all_keywords	0.7662	0.0120
Comp2023.stack_3_bow_tailored_all_keywords	0.7661	0.0200
Comp2023.stack_3_bows_tailored_keywords	0.7640	0.0120
Comp2023.stack_3_bows	0.7575	0.0020
Comp2023.bow_training_set	0.7379	0.0020
Comp2023.bow	0.7300	0.0000
Comp2023.bow_voc_selection	0.7221	0.0000