A5025269577- Topic Modeling
- Natural Language Processing Techniques
- Advanced Text Analysis Techniques
- Sentiment Analysis and Opinion Mining
- Semantic Web and Ontologies
- Text and Document Classification Technologies
- Authorship Attribution and Profiling
- Complex Network Analysis Techniques
- Speech and dialogue systems
- Text Readability and Simplification
- Web Data Mining and Analysis
- Misinformation and Its Impacts
- Seismology and Earthquake Studies
- Computational and Text Analysis Methods
- Spam and Phishing Detection
- Multimodal Machine Learning Applications
- Wine Industry and Tourism
- Disaster Management and Resilience
- Open Source Software Innovations
- EEG and Brain-Computer Interfaces
- Public Relations and Crisis Communication
- Research Data Management Practices
- Technology and Data Analysis
- Algorithms and Data Compression
- Web visibility and informetrics
University of West Bohemia
2013-2024
National Centre of Scientific Research "Demokritos"
2017
University of Alicante
2017
Sami Shamoon College of Engineering
2017
University of Bari Aldo Moro
2017
Software (Spain)
2017
Joint Research Centre
2010-2012
Joint Research Centre
2012
This paper describes our system participating in the aspect-based sentiment analysis task of Semeval 2014.The goal was to identify aspects given target entities and expressed towards each aspect.We firstly introduce a based on supervised machine learning, which is strictly constrained uses training data as only source information.This then extended by unsupervised methods for latent semantics discovery (LDA semantic spaces) well approach vocabularies.The evaluation done two domains,...
This paper describes the outcomes of first challenge on multilingual named entity recognition that aimed at recognizing mentions entities in web documents Slavic languages, their normalization/lemmatization, and cross-language matching. It was organised context 6th Balto-Slavic Natural Language Processing Workshop, co-located with EACL 2017 conference. Although eleven teams signed up for evaluation, due to complexity task(s) short time available elaborating a solution, only two submitted...
George Giannakopoulos, Jeff Kubina, John Conroy, Josef Steinberger, Benoit Favre, Mijail Kabadjov, Udo Kruschwitz, Massimo Poesio. Proceedings of the 16th Annual Meeting Special Interest Group on Discourse and Dialogue. 2015.
Jakub Piskorski, Laska Laskova, Michał Marcińczuk, Lidia Pivovarova, Pavel Přibáň, Josef Steinberger, Roman Yangarber. Proceedings of the 7th Workshop on Balto-Slavic Natural Language Processing. 2019.
This paper presents a pioneering research on aspect-level sentiment analysis in Czech. The main contribution of the is newly created Czech aspectlevel corpus, based data from restaurant reviews. We annotated corpus with two variants ‐ aspect terms and categories. consists 1,244 sentences 1,824 aspects freely available to community. Furthermore, we propose baseline system supervised machine learning. Our detects Fmeasure 68.65% their polarities accuracy 66.27%. categories are recognized...
We examine the effectiveness of several unsupervised methods for latent semantics discovery as features aspect-based sentiment analysis (ABSA). use shared task definition from SemEval 2014. In our experiments we labeled and unlabeled corpora within restaurants domain two languages: Czech English. show that models improve ABSA performance prove approach is worth exploring. Moreover, achieve new state-of-the-art results Czech. Another important contribution work created restaurant task: one...
This paper describes our system created to detect stance in online discussions. The goal is identify whether the author of a comment favor given target or against. Our approach based on maximum entropy classifier, which uses surface-level, sentiment and domain-specific features. was originally developed English tweets. We adapted it process Czech news commentaries.
This paper describes our system participating in the SemEval 2016 task: Detecting stance Tweets.The goal was to identify whether author of a tweet is favor given target or against.Our approach based on maximum entropy classifier, which uses surface-level, sentiment and domain-specific features.We participated both supervised weakly subtasks received promising results for most targets.
We propose an approach to summarization exploiting both lexical information and the output of automatic anaphoric resolver, using Singular Value Decomposition (SVD) identify main terms. demonstrate that adding results in significant performance improvements over a previously developed system, which only terms are used as input SVD. However, we also show how is crucial: whereas this add new does result improved performance, simple substitution makes worse.
This paper deals with our recent research in text summarization. The field has moved from multi-document summarization to update When producing an summary of a set topic-related documents the summarizer assumes prior knowledge reader determined by older same topic. thus must solve novelty vs. redundancy problem. We describe development which is based on Iterative Residual Rescaling (IRR) that creates latent semantic space under consideration. IRR generalizes Singular Value Decomposition...