Description of corpora
Corpora were constituted from two separate sources: debates in the
European Parliament
and a collection of articles from the newspapers
Le Monde (France), The Financial Times (United Kingdom) and Il Sole 24 Ore (Italia).
This corpus contains 32,289 interventions by representatives,
which took place at the Parliament between 1999 and 2004.
We only extracted interventions from representatives belonging to one of
following five parties:
- Verts-ALE : The Greens/European Free Alliance;
- GUE-NGL : Confederal Group of the European United Left/Nordic Green Left;
- PSE : Party of European Socialists;
- ELDR : European Liberal Democrat and Reform Party;
- PPE-DE : European People's Party (Christian Democrats) and European Democrats
Examples
- Party: Verts-ALE
- Mr President, yesterday once again eleven people from Africa died off the Spanish coast when their boat sank. We cannot remain indifferent to this tragedy, which occurs very often: although it will be difficult, we must seek a so
lution so that these people do not feel obliged to try to reach European territory in such a dramatic, and so often tragic, way.
I believe it is honourable to remember this now, under these circumstances.
- Party: GUE-NGL
- Madam President, the Commission's proposal for an amendment to the common organisation of the market in fruit and vegetables is exacerbating the problems of the current market organisation and the unfairness of the CAP. It is also creating greater difficulties for fruit and vegetable growers. Particularly serious are the measures designed to abolish the minimum price, which affects tomatoes destined for processing, the reduction in the limit for aid for operational funds from 4.5% to 3% of the value of the production marketed by each producer organisation, the precautionary reduction of 9.1% of the amount of aid for the first marketing year following the reform of the COM and a reduction in the citrus fruit withdrawal ceiling.
- Party: PSE
- Mr President, ladies and gentlemen, I would like to confine my remarks to the subject of HIV/AIDS. The World Population Report for 2002, which has just appeared, contains some shocking figures. On average, 14 000 men, women and children die of this scourge every day. It is now the commonest cause of death in Sub-Saharan Africa. To date, more than 60 million people worldwide have become infected and around 22 million of them have died. Of the 40 million persons infected, 95% live in developing countries and nearly three quarters of them in Africa. Out of 580 000 children below the age of 15 who have died from AIDS, 500 000 - almost 90% - lived in Africa. I could continue with these statistics of misery indefinitely.
- Party: ELDR
- Mr President, the EU is the key donor to the Palestinian territories. According to the UN, since the closures, more than one million Palestinians have been living under the poverty line of two dollars a day. The human effect of the closures is possibly even worse. Commissioner Patten stated it a moment ago: sick people cannot even get to the European hospital in Gaza because even ambulances cannot get through the barricades. It is therefore of key importance that we continue providing aid. I am delighted that the Council and Commission share our view.
- Party: PPE-DE
- Madam President, it was with great regret that I had to vote against the budget as I find the items devoted to improving the living conditions of the elderly and pensioners to be totally inadequate, or rather completely non-existent. I furthermore note that a significant proportion of these funds have been earmarked for the notorious Community action programmes. It is my opinion that these programmes fail to perform the beneficial role that they should in terms of the utilisation of Community funds. I believe that the European Union must completely reassess the way in which it spends the monies of the 15 Member States.
The assignment of « objective » and « subjective » values
to articles was made in different ways for different newspapers.
Le Monde : a French newspaper of general interest, corpus of 42,000 articles
for the period 2003/2006 ;
The Financial Times : a British economic newspaper, corpus of 13,000 articles for the year 1993 ;
Il Sole 24 Ore : an Italian economic newspaper, corpus of 2,500 articles for the year 1992.
Examples "Le Monde"
- Property: objective
- SOUPÇONNÉ par la direction de la surveillance du territoire
(DST) d'être l'un des informateurs anonymes des juges Renaud Van
Ruymbeke et Dominique de Talancé, qui enquêtent sur l'affaire des
frégates de Taïwan, Imad Lahoud, informaticien chez EADS, rompt le
silence par la voix de son avocat. Dans un communiqué adressé au
Monde, Me Olivier Pardo assure qu' « en dépit de la multiplicité des
assertions aucune preuve d'une quelconque participation de -son-
client à cette affaire n'existe ». « Dans une ronde sans fin, la
calomnie s'installe et risque de devenir extrêmement préjudiciable
», ajoute Me Pardo, qui précise que son client « n'a jamais
rencontré le député Alain Marsaud », contrairement à ce qu'a laissé
entendre la DST dans un rapport ( Le Monde du 20 juillet). «
M. Lahoud demande que cessent ces assertions et
l'instrumentalisation de sa personne, conclut l'avocat. Il ne
réclame qu'une chose : qu'il puisse continuer à animer son équipe de
recherche dans la grande entreprise européenne à laquelle il a
l'honneur d'appartenir, dans la sérénité et la quiétude. »
- Property: subjective
- Avec la campagne présidentielle qui se profile, voici revenu
le temps des effets chocs et des idées chics. Il en est ainsi du «
dialogue social », serpent de mer du débat public que les politiques
défendent surtout quand ils n'ont pas à le pratiquer. Jacques Chirac
en a parlé le 14 juillet. Dominique de Villepin redécouvre son
existence après l'avoir superbement ignoré. L'UMP va achever, en
septembre, des rencontres avec tous les partenaires sociaux, CGT
comprise. Le PS recevra, à son Université d'été de La Rochelle, du
25 au 27 août, tous les syndicats et le numéro deux du Medef, Denis
Gautier-Sauvagnac.
Examples "The Financial Times"
- Property: objective
- THE Health and Safety Executive yesterday accused the
European Community of imposing a mass of health and safety legislation
on member states without adequate thought or consultation, Diane
Summers writes.
The executive is concerned that it is becoming a
focus of complaints that UK businesses are increasingly subject to
excessively bureaucratic regulation. It is anxious to remind the
government that much of the regulation and perceived red tape
originates from Brussels and not from it.
Mr John Rimington, the
executive's director-general, said the directives had been constructed
in 'smoke-filled rooms in Brussels'. He blamed, in particular, the
French for trying to get the directives to reflect their own domestic
laws and said some of the provisions were 'incomprehensible'.
- Property: subjective
- Denmark has a new government. The foreign minister has
pledged himself to secure a Yes to Maastricht in the second Danish
referendum on the subject, to be held probably 'before June'. All's
right with the world.
Some moaning Euro-minnies are still
muttering about the dangerous precedent set by Denmark's special
'opt-outs', negotiated at last month's Edinburgh summit. Won't
Conservative backbenchers try to obtain the same deal for Britain, as
the price of ratification? Won't candidates for EC membership, with
three of whom formal negotiations are to start on Monday, demand that
the same exemptions apply to them? Isn't this the beginning of the a
la carte union so dreaded by Mr Jacques Delors, president of the
European Commission?
Examples "Il Sole 24 Ore"
- Property: objective
- Sarà la presidenza del Consiglio a dire l'ultima parola sulla
riforma dei fondi pensione. Il ministro del Lavoro Nino Cristofori
ha già annunciato per il mese di dicembre la presentazione del suo
progetto di regolamentazione delle casse integrative aziendali, in
attuazione della recente legge delega sul riordino della
previdenza. Quello di Cristofori, però, sarà soltanto il materiale
preparatorio per la stesura definitiva del decreto legislativo che
darà il via libera alla riforma. A Palazzo Chigi si è costituito in
questi giorni un gruppo di esperti con il compito di analizzare le
proposte del Lavoro confrontandole con quelle provenienti da altri
ministeri (Tesoro, Industria e Finanze), parti sociali e gruppi
economici.
- Property: subjective
- L'economia va male e quindi i tassi di interesse si
riducono. Questa apparentemente banale relazione è tornata a essere
vera in questi giorni un pò in tutta Europa e quindi anche in
Italia. Dopo i fuochi d'artificio di un mese fa, quando sembrava che
le banche centrali di tanti Paesi europei fossero solo impegnate
nella nobile gara a chi alzava di più i tassi di interesse, il
buonsenso economico è tornato a prevalere.
The corpora are in an XML format with a DTD available here
(last update: 01/12/08).
Lists of values for article properties (Task 1) and political parties (Task 3) are folllowing:
- « Newspaper » Corpus: OBJECTIF, SUBJECTIF.
- « Parliament » Corpus: Verts-ALE, GUE-NGL, PSE, ELDR, PPE-DE ;
For Tasks 1 and 3, reference files are available, therefore,
learning processes can be developed. Contrariwise, for Task 2,
there are no reference files.
Examples
- learning files:
- « Parliament » Corpus:
each document contains the text of the parliamentary intervention
and the name of the party the representative belongs to.
(examples in French,
English and
Italian).
- « Newspaper » Corpus:
each document contains a newspaper article
and the subjective or objective nature of this article
(examples from Le Monde,
The Financial Times and
Il Sole 24 Ore).
- Test files: each file contains, at the document level, only a text
(parliamentary intervention or newspaper article);
- Result files (see also a detailed description):
- Task 1: the file must contain the evaluation of the objective/subjective property,
without the text of the document
(examples in French,
English and
Italian)
- Task 2: the file must contain, for each document, its text
with the <SUBJECTIF> and </SUBJECTIF> tags for
fragments identified as subjective (examples in
French,
English and
Italian)
- Task 3: the file must contain, for each document, the evaluation of
the speaker's party, without the text of the document
(examples in French,
English and
Italian)
The encoding of corpora was modified to be uniformly UTF-8.
Possible misspellings or mispunctuations were not corrected.
The userid and password requested for accessing the corpora
will be sent to persons who will have registrated
and who will have signed and sent the agreement.
Training Corpus
Update: 09/01/21.
Test Corpus
Update: 09/03/18.
Reference data
➪ F-measure perl scripts: task 1 and 3
Update: 09/05/05.
