The PHASAR Search Engine

Koster, C. H. A.; Seibert, Olaf; Seutter, Marc

doi:10.1007/11765448_13

Cited by 5 publications

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the TM4IP project, a text mining system is being implemented for intellectual property search [11]. The system consists of (1) an English hybrid dependency parser (Aegir) that is especially developed to deal with the syntactic characteristics and vocabulary of technical texts, more specifically patent texts, and (2) a professional interactive search engine that uses structured queries based on dependency relations to find patent documents in a database [12]. The Aegir parser combines a set of handwritten syntactic rules with an extensive word form lexicon ('the parser lexicon') and information about the frequencies of dependency relations between words.…”

Section: The Aegir Parser In Tm4ipmentioning

confidence: 99%

Genre and domain in patent texts

Oostdijk

D’Hondt

Halteren

et al. 2010

Proceedings of the 3rd International Workshop on Patent Information Retrieval

View full text Add to dashboard Cite

In this paper we investigate the variation in language use within the very broad patent domain. We find that language use (represented by syntactic phrases) not only differs from one patent class to the next, but is also a characteristic that sets apart the four sections of a patent (viz. Title, Abstract, Description and Claims). This lends support to the claim that these sections can be viewed as different text genres. For the development of a syntactic parser that is trained on patent texts, we quantify the domain and genre differences in terms of the amounts of text needed to train domain-dependent versions of the parser. Our quantified and exemplified findings on the domain variation in patent data are of interest for the patent retrieval and analysis communities. KeywordsParsing, Patent domains, Text mining 1 A e g ir stands for Accurate English Grammar for Information Retrieval.

show abstract

Section: The Aegir Parser In Tm4ipmentioning

confidence: 99%

Genre and domain in patent texts

Oostdijk

D’Hondt

Halteren

et al. 2010

Proceedings of the 3rd International Workshop on Patent Information Retrieval

View full text Add to dashboard Cite

show abstract

“…There are, therefore, many different approaches for information extraction in the biomedical text. Some applications are geared towards helping with automatic population of interaction databases [ 7 , 8 ], while others aim to support a wide variety of users by bridging the gap between the search engines and highly customised relation extraction software [ 9 - 12 ]. Different approaches to PPI detection can be roughly categorised into pattern-based, information retrieval-based (IR-based), and classification-based [ 7 , 13 - 15 ].…”

Section: Introductionmentioning

confidence: 99%

“…A favourite approach of initial systems aiming to construct interaction networks on the fly from user queries, it is an efficient way of allowing the user to browse potential interactions [ 9 , 10 , 18 ]. More advanced IR-based approaches incorporate interaction detection into the indexing process [ 11 , 12 ]. This allows for fast retrieval of highly detailed information.…”

Section: Introductionmentioning

confidence: 99%

Protein interaction sentence detection using multiple semantic kernels

2011

View full text Add to dashboard Cite

BackgroundDetection of sentences that describe protein-protein interactions (PPIs) in biomedical publications is a challenging and unresolved pattern recognition problem. Many state-of-the-art approaches for this task employ kernel classification methods, in particular support vector machines (SVMs). In this work we propose a novel data integration approach that utilises semantic kernels and a kernel classification method that is a probabilistic analogue to SVMs. Semantic kernels are created from statistical information gathered from large amounts of unlabelled text using lexical semantic models. Several semantic kernels are then fused into an overall composite classification space. In this initial study, we use simple features in order to examine whether the use of combinations of kernels constructed using word-based semantic models can improve PPI sentence detection.ResultsWe show that combinations of semantic kernels lead to statistically significant improvements in recognition rates and receiver operating characteristic (ROC) scores over the plain Gaussian kernel, when applied to a well-known labelled collection of abstracts. The proposed kernel composition method also allows us to automatically infer the most discriminative kernels.ConclusionsThe results from this paper indicate that using semantic information from unlabelled text, and combinations of such information, can be valuable for classification of short texts such as PPI sentences. This study, however, is only a first step in evaluation of semantic kernels and probabilistic multiple kernel learning in the context of PPI detection. The method described herein is modular, and can be applied with a variety of feature types, kernels, and semantic models, in order to facilitate full extraction of interacting proteins.

show abstract

“…Finally, the words in the DT's are lemmatized. The EP4IR parser/transducer was developed for application in Information Retrieval [11]. Our paper shows that it can also be used successfully for Domain Modeling.…”

Section: Dependency Trees and Dependency Tripletsmentioning

confidence: 99%

Automatic Issue Extraction from a Focused Dialogue

Hindriks

Hoppenbrouwers

Jonker

et al.

Natural Language Processing and Information Systems

View full text Add to dashboard Cite

Abstract.Various methodologies for structuring the process of domain modeling have been proposed, but there are few software tools that provide automatic support for the process of constructing a domain model. The problem is that it is hard to extract the relevant concepts from natural language texts since these typically include many irrelevant details that are hard to discern from relevant concepts. In this paper, we propose an alternative approach to extract domain models from natural language input. The idea is that more effective, automatic extraction is possible from a natural language text that is produced in a focused dialogue game. We present an application of this idea in the area of pre-negotiation, in combination with sophisticated parsing and transduction techniques for natural language and fairly simple pattern matching rules. Furthermore, a prototype is presented of a conversation-oriented experimentation environment for cooperative conceptualization. Several experiments have been performed to evaluate the approach and environment, and a technique for measuring the quality of extraction has been defined. The experiments indicate that even with a simple implementation of the proposed approach reasonably acceptable results can be obtained.

show abstract

The PHASAR Search Engine

Cited by 5 publications

References 10 publications

Genre and domain in patent texts

Genre and domain in patent texts

Protein interaction sentence detection using multiple semantic kernels

Automatic Issue Extraction from a Focused Dialogue

Contact Info

Product

Resources

About