COATI: Multilingual advanced search in blogs for recovery of opinion and trends in the business and public administration fields
The exponential increase of websites has entailed the increasingly difficult for companies and governments to track the actual impact of their products, policies, and activities; to know the public opinion or to anticipate emerging trends on the web. With the rise of blogs and social media appeared a new phenomenon: opinions and trends in different formats and languages, meaning new technological challenges in the field of information retrieval.
Opinion mining is a new and emerging technology necessary in the globalized society of the web 2.0. This discipline is a recent result of crossing information retrieval, computational linguistics and artificial intelligence, which works not only with the information in a document but also with the views collected.
The project COATI Opinion Mining develops a platform of opinions and trends in specialized blogs from which companies and governments can get so accurate relevant information about products, projects and policies.
Objectives
- Carry out a technical study on the challenges in access and use of information in blogs and reference collections of blogs and issues of crawling (diversity of sources, formats, refresh rates).
- Orientation to profiles: oriented platform to user characteristics and complex information needs.
- Multilingualism: the user should only need to express their opinions and trends in his language. The system will search information in different languages (Galician, Spanish, Portuguese, French and English) and return results in different languages using the translation system Opentrad.
- Relevant results oriented: there will be used text summarization techniques to condense the information as much as possible and minimize the time required for processing.
- Open source: the resulting platform will be released following the recommendations of the group European ISTAG under GPL from the Free Software Foundation so it can be exploited for future developments made by the community.
- Construction of a multilingual opinion mining prototype for application in the field of journalism, telephone, interactive TV and in public administration.
Project
/research/projects/procura-avanzada-multilingue-en-blogs-para-a-recuperacion-de-opinion-e-tendencias-nos-ambitos-empresarial-e-da-administracion-publica
<p> The exponential increase of websites has entailed the increasingly difficult for companies and governments to track the actual impact of their products, policies, and activities; to know the public opinion or to anticipate emerging trends on the web. With the rise of blogs and social media appeared a new phenomenon: opinions and trends in different formats and languages, meaning new technological challenges in the field of information retrieval. </p><p>Opinion mining is a new and emerging technology necessary in the globalized society of the web 2.0. This discipline is a recent result of crossing information retrieval, computational linguistics and artificial intelligence, which works not only with the information in a document but also with the views collected. </p><p> The project COATI Opinion Mining develops a platform of opinions and trends in specialized blogs from which companies and governments can get so accurate relevant information about products, projects and policies. </p><ul><li>Carry out a technical study on the challenges in access and use of information in blogs and reference collections of blogs and issues of crawling (diversity of sources, formats, refresh rates).</li><li> Orientation to profiles: oriented platform to user characteristics and complex information needs.</li><li> Multilingualism: the user should only need to express their opinions and trends in his language. The system will search information in different languages (Galician, Spanish, Portuguese, French and English) and return results in different languages using the translation system Opentrad. </ li> <li> Relevant results oriented: there will be used text summarization techniques to condense the information as much as possible and minimize the time required for processing.</li><li>Open source: the resulting platform will be released following the recommendations of the group European ISTAG under GPL from the Free Software Foundation so it can be exploited for future developments made by the community.</li><li>Construction of a multilingual opinion mining prototype for application in the field of journalism, telephone, interactive TV and in public administration. </ li> </ ul> - David Enrique Losada Carril
projects_en