Combining sequence and itemset mining to discover named entities in biomedical texts: a new type of pattern

Plantevit, Marc; Charnois, Thierry; Kléma, Jǐŕı; Rigotti, Christophe; Crémilleux, Bruno

doi:10.1504/ijdmmm.2009.026073

Cited by 11 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…But there are also applications of FSM for textual data. Plantevit et al [13] present a method for FSM for Biomedical named entity recognition task.…”

Section: Related Workmentioning

confidence: 99%

Clinical Text Mining for Context Sequences Identification

Boytcheva

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper presents an approach based on sequence mining for identification of context models of diseases described by different medical specialists in clinical text. Clinical narratives contain rich medical terminology, specific abbreviations, and various numerical values. Usually raw clinical texts contain too many typos. Due to the telegraphic style of the text and incomplete sentences, the general part of speech taggers and syntax parsers are not efficient in text processing of non-English clinical text. The proposed approach is language independent. Thus, the method is suitable for processing clinical texts in low resource languages. The experiments are done on pseudonimized outpatient records in Bulgarian language produced by four different specialists for the same cohort of patients suffering from similar disorders. The results show that from the clinical documents can be identified the specialty of the physician. Even the close vocabulary is used in the patient status description there are slight differences in the language used by different physicians. The depth and the details of the description allow to determine different aspects and to identify the focus in the text. The proposed data driven approach will help for automatic clinical text classification depending on the specialty of the physician who wrote the document. The experimental results show high precision and recall in classification task for all classes of specialist represented in the dataset. The comparison of the proposed method with bag of words method show some improvement of the results in document classification task.

show abstract

“…But there are also applications of FSM for textual data. Plantevit et al [13] present a method for FSM for Biomedical named entity recognition task.…”

Section: Related Workmentioning

confidence: 99%

Clinical Text Mining for Context Sequences Identification

Boytcheva

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Association rule and sequence mining approaches have been similarly used to identify relationships between constructs in a text [for example, 1,20,35], and to detect erroneous sentences [42]. Linguistics research has used these approaches to investigate the ways that language is structured, and used in everyday contexts.…”

Section: Sequence and Process Analyses Of Student Writingmentioning

confidence: 99%

Towards mining sequences and dispersion of rhetorical moves in student written texts

Knight

Martínez-Maldonado

Gibson

et al. 2017

Proceedings of the Seventh International Learning Analytics &Amp; Knowledge Conference

View full text Add to dashboard Cite

There is an increasing interest in the analysis of both student's writing and the temporal aspects of learning data. The analysis of higher-level learning features in writing contexts requires analyses of data that could be characterised in terms of the sequences and processes of textual features present. This paper (1) discusses the extant literature on sequential and process analyses of writing; and, based on this and our own first-hand experience on sequential analysis, (2) proposes a number of approaches to both pre-process and analyse sequences in whole-texts. We illustrate how the approaches could be applied to examples drawn from our own datasets of 'rhetorical moves' in written texts, and the potential each approach holds for providing insight into that data. Work is in progress to apply this model to provide empirical insights. Although, similar sequence or process mining techniques have not yet been applied to student writing, techniques applied to event data could readily be operationalised to undercover patterns in texts.

show abstract

“…Plantevit et al [28] developed left-sequence-right (LSR) patterns to search for named entities from real datasets BioCreative, Genia and Abstracts, taking into account the surrounding context of a sequence and relaxing the order constraint around the sequence. The presence of sequential pattern mining algorithms in phrase extraction from massive text corpora by Liu et al [12] and Ren et al [13] in 2015 will certainly promote more research on medical term extraction to adopt such an approach.…”

Section: Related Workmentioning

confidence: 99%

A genetic algorithm enabled ensemble for unsupervised medical term extraction from clinical letters

Liu

Chung

Wang

et al. 2015

Health Inf Sci Syst

View full text Add to dashboard Cite

Despite the rapid global movement towards electronic health records, clinical letters written in unstructured natural languages are still the preferred form of inter-practitioner communication about patients. These letters, when archived over a long period of time, provide invaluable longitudinal clinical details on individual and populations of patients. In this paper we present three unsupervised approaches, sequential pattern mining (PrefixSpan); frequency linguistic based C-Value; and keyphrase extraction from co-occurrence graphs (TextRank), to automatically extract single and multi-word medical terms without domain-specific knowledge. Because each of the three approaches focuses on different aspects of the language feature space, we propose a genetic algorithm to learn the best parameters of linearly integrating the three extractors for optimal performance against domain expert annotations. Around 30,000 clinical letters sent over the past decade from ophthalmology specialists to general practitioners at an eye clinic are anonymised as the corpus to evaluate the effectiveness of the ensemble against individual extractors. With minimal annotation, the ensemble achieves an average F-measure of 65.65 % when considering only complex medical terms, and a F-measure of 72.47 % if we take single word terms (i.e. unigrams) into consideration, markedly better than the three term extraction techniques when used alone.

show abstract

Combining sequence and itemset mining to discover named entities in biomedical texts: a new type of pattern

Cited by 11 publications

References 30 publications

Clinical Text Mining for Context Sequences Identification

Clinical Text Mining for Context Sequences Identification

Towards mining sequences and dispersion of rhetorical moves in student written texts

A genetic algorithm enabled ensemble for unsupervised medical term extraction from clinical letters

Contact Info

Product

Resources

About