http://hlt.fbk.eu/corps
Description
CORPS is a corpus of political speeches tagged with specific audience reactions, such as APPLAUSE or LAUGHTER.
In collecting this corpus we relied on the hypothesis that such tags about public reaction are indicators of hotspots, where persuasion attempts succeeded or, at least, a persuasive attempt has been recognized by the audience. This corpus can be usefully employed in many fields such as:
qualitative analysis of political communication.
NLP based persuasive expression mining.
Automatic production of persuasive communication
At present, there are about 900 speeches in the corpus and about 2.2 millions words.
The speeches are all native English language, and all represent monological situations (i.e. there is only one speaker addressing an audience).
These speeches have been collected from internet, and an automatic conversion of audience reactions tags - to make them homogeneous in formalism and labeling - has been performed.
Metadata regarding the speech has also been added (title, event, speaker, date, description).
How to obtain it
CORPS is freely available for research purposes, for further information please write to guerini[at]fbk.eu and strappa[at]fbk.eu
Description
CORPS is a corpus of political speeches tagged with specific audience reactions, such as APPLAUSE or LAUGHTER.
In collecting this corpus we relied on the hypothesis that such tags about public reaction are indicators of hotspots, where persuasion attempts succeeded or, at least, a persuasive attempt has been recognized by the audience. This corpus can be usefully employed in many fields such as:
qualitative analysis of political communication.
NLP based persuasive expression mining.
Automatic production of persuasive communication
At present, there are about 900 speeches in the corpus and about 2.2 millions words.
The speeches are all native English language, and all represent monological situations (i.e. there is only one speaker addressing an audience).
These speeches have been collected from internet, and an automatic conversion of audience reactions tags - to make them homogeneous in formalism and labeling - has been performed.
Metadata regarding the speech has also been added (title, event, speaker, date, description).
How to obtain it
CORPS is freely available for research purposes, for further information please write to guerini[at]fbk.eu and strappa[at]fbk.eu