In this paper, a robust parser for Swedish is presented. The parser identifies the internal structure of phrases, but does not build full trees. In addition to phrase identification, clause boundaries are detected. The parser is designed for robustness against noisy and ill-formed data. An evaluation on 15 000 words shows that the parser’s accuracy on phrase bracketing is 88.7 per cent and the F-score for clause boundary identification is 88.3 per cent. 1

In this paper, we describe an empirical study of Chinese chunking on a corpus, which is extracted from UPENN Chinese Treebank-4 (CTB4). First, we compare the performance of the state-of-the-art machine learning models. Then we propose two approaches in order to improve the performance of Chinese chunking. 1) We propose an approach to resolve the special problems of Chinese chunking. This approach extends the chunk tags for every problem by a tag-extension function. 2) We propose two novel voting methods based on the characteristics of chunking task. Compared with traditional voting methods, the proposed voting methods consider long distance information. The experimental results show that the SVMs model outperforms the other models and that our proposed approaches can improve performance significantly. 1

Abstract. The presented paper describes a method of text preprocessing improving the performance of sequential data mining applied in the task of gene interaction extraction from biomedical texts. The need of text preprocessing rises primarily from the fact, that the language encoded by any general word sequence is mostly not sequential. The method involves a number of heuristic language transformations, all together converting sentences into forms with higher degree of sequentiality. The core idea of enhancing sentence sequentiality results from the observation that the components constituting the semantical and grammatical content of sentences are not equally relevant for extracting a highly specific type of information. Experiments employing a simple sequential algorithm confirmed the usability of the proposed text preprocessing in the gene interaction extraction task. Furthermore, limitations identified during the result analysis may be regarded as guidelines for further work exploring the capabilities of the sequential data mining applied on linguistically preprocessed texts.