Character-level neural network for biomedical named entity recognition

Biomedical named entity recognition (BNER), which extracts important named entities such as genes and proteins, is a challenging task in automated systems that mine knowledge in biomedical texts. The previous state-of-the-art systems required large amounts of task-specific knowledge in the form of feature engineering, lexicons and data pre-processing to achieve high performance. In this paper, we introduce a novel neural network architecture that benefits from both word- and character-level representations automatically, by using a combination of bidirectional long short-term memory (LSTM) and conditional random field (CRF) eliminating the need for most feature engineering tasks. We evaluate our system on two datasets: JNLPBA corpus and the BioCreAtIvE II Gene Mention (GM) corpus. We obtained state-of-the-art performance by outperforming the previous systems. To the best of our knowledge, we are the first to investigate the combination of deep neural networks, CRF, word embeddings and character-level representation in recognizing biomedical named entities.     

Enhancing HMM-based biomedical named entity recognition by studying special phenomena

The purpose of this research is to enhance an HMM-based named entity recognizer in the biomedical domain. First, we analyze the characteristics of biomedical named entities. Then, we propose a rich set of features, including orthographic, morphological, part-of-speech, and semantic trigger features. All these features are integrated via a Hidden Markov Model with back-off modeling. Furthermore, we propose a method for biomedical abbreviation recognition and two methods for cascaded named entity recognition. Evaluation on the GENIA V3.02 and V1.1 shows that our system achieves 66.5 and 62.5 F-measure, respectively, and outperforms the previous best published system by 8.1 F-measure on the same experimental setting. The major contribution of this paper lies in its rich feature set specially designed for biomedical domain and the effective methods for abbreviation and cascaded named entity recognition. To our best knowledge, our system is the first one that copes with the cascaded phenomena.     

Stacked ensemble combined with fuzzy matching for biomedical named entity recognition of diseases

Biomedical Named Entity Recognition (Bio-NER) is the crucial initial step in the information extraction process and a majorly focused research area in biomedical text mining. In the past years, several models and methodologies have been proposed for the recognition of semantic types related to gene, protein, chemical, drug and other biological relevant named entities. In this paper, we implemented a stacked ensemble approach combined with fuzzy matching for biomedical named entity recognition of disease names. The underlying concept of stacked generalization is to combine the outputs of base-level classifiers using a second-level meta-classifier in an ensemble. We used Conditional Random Field (CRF) as the underlying classification method that makes use of a diverse set of features, mostly based on domain specific, and are orthographic and morphologically relevant. In addition, we used fuzzy string matching to tag rare disease names from our in-house disease dictionary. For fuzzy matching, we incorporated two best fuzzy search algorithms Rabin Karp and Tuned Boyer Moore. Our proposed approach shows promised result of 94.66%, 89.12%, 84.10%, and 76.71% of F-measure while on evaluating training and testing set of both NCBI disease and BioCreative V CDR Corpora.     

Unsupervised biomedical named entity recognition: Experiments with clinical and biological texts

Named entity recognition is a crucial component of biomedical natural language processing, enabling information extraction and ultimately reasoning over and knowledge discovery from text. Much progress has been made in the design of rule-based and supervised tools, but they are often genre and task dependent. As such, adapting them to different genres of text or identifying new types of entities requires major effort in re-annotation or rule development. In this paper, we propose an unsupervised approach to extracting named entities from biomedical text. We describe a stepwise solution to tackle the challenges of entity boundary detection and entity type classification without relying on any handcrafted rules, heuristics, or annotated data. A noun phrase chunker followed by a filter based on inverse document frequency extracts candidate entities from free text. Classification of candidate entities into categories of interest is carried out by leveraging principles from distributional semantics. Experiments show that our system, especially the entity classification step, yields competitive results on two popular biomedical datasets of clinical notes and biological literature, and outperforms a baseline dictionary match approach. Detailed error analysis provides a road map for future work.     

Automatic recognition of disorders,findings, pharmaceuticals and body structures from clinical text: An annotation and machine learning study

Automatic recognition of clinical entities in the narrative text of health records is useful for constructing applications for documentation of patient care, as well as for secondary usage in the form of medical knowledge extraction. There are a number of named entity recognition studies on English clinical text, but less work has been carried out on clinical text in other languages.This study was performed on Swedish health records, and focused on four entities that are highly relevant for constructing a patient overview and for medical hypothesis generation, namely the entities: Disorder, Finding, Pharmaceutical Drug and Body Structure. The study had two aims: to explore how well named entity recognition methods previously applied to English clinical text perform on similar texts written in Swedish; and to evaluate whether it is meaningful to divide the more general category Medical Problem, which has been used in a number of previous studies, into the two more granular entities, Disorder and Finding.Clinical notes from a Swedish internal medicine emergency unit were annotated for the four selected entity categories, and the inter-annotator agreement between two pairs of annotators was measured, resulting in an average F-score of 0.79 for Disorder, 0.66 for Finding, 0.90 for Pharmaceutical Drug and 0.80 for Body Structure. A subset of the developed corpus was thereafter used for finding suitable features for training a conditional random fields model. Finally, a new model was trained on this subset, using the best features and settings, and its ability to generalise to held-out data was evaluated. This final model obtained an F-score of 0.81 for Disorder, 0.69 for Finding, 0.88 for Pharmaceutical Drug, 0.85 for Body Structure and 0.78 for the combined category Disorder + Finding.The obtained results, which are in line with or slightly lower than those for similar studies on English clinical text, many of them conducted using a larger training data set, show that the approaches used for English are also suitable for Swedish clinical text. However, a small proportion of the errors made by the model are less likely to occur in English text, showing that results might be improved by further tailoring the system to clinical Swedish. The entity recognition results for the individual entities Disorder and Finding show that it is meaningful to separate the general category Medical Problem into these two more granular entity types, e.g. for knowledge mining of co-morbidity relations and disorder-finding relations.     

Semi-supervised medical entity recognition: A study on Spanish and Swedish clinical corpora

ObjectiveThe goal of this study is to investigate entity recognition within Electronic Health Records (EHRs) focusing on Spanish and Swedish. Of particular importance is a robust representation of the entities. In our case, we utilized unsupervised methods to generate such representations.MethodsThe significance of this work stands on its experimental layout. The experiments were carried out under the same conditions for both languages. Several classification approaches were explored: maximum probability, CRF, Perceptron and SVM. The classifiers were enhanced by means of ensembles of semantic spaces and ensembles of Brown trees. In order to mitigate sparsity of data, without a significant increase in the dimension of the decision space, we propose the use of clustered approaches of the hierarchical Brown clustering represented by trees and vector quantization for each semantic space.ResultsThe results showed that the semi-supervised approaches significantly improved standard supervised techniques for both languages. Moreover, clustering the semantic spaces contributed to the quality of the entity recognition while keeping the dimension of the feature-space two orders of magnitude lower than when directly using the semantic spaces.ConclusionsThe contributions of this study are: (a) a set of thorough experiments that enable comparisons regarding the influence of different types of features on different classifiers, exploring two languages other than English; and (b) the use of ensembles of clusters of Brown trees and semantic spaces on EHRs to tackle the problem of scarcity of available annotated data.     

A controlled greedy supervised approach for co-reference resolution on clinical text

Identification of co-referent entity mentions inside text has significant importance for other natural language processing (NLP) tasks (e.g. event linking). However, this task, known as co-reference resolution, remains a complex problem, partly because of the confusion over different evaluation metrics and partly because the well-researched existing methodologies do not perform well on new domains such as clinical records. This paper presents a variant of the influential mention-pair model for co-reference resolution. Using a series of linguistically and semantically motivated constraints, the proposed approach controls generation of less-informative/sub-optimal training and test instances. Additionally, the approach also introduces some aggressive greedy strategies in chain clustering. The proposed approach has been tested on the official test corpus of the recently held i2b2/VA 2011 challenge. It achieves an unweighted average F₁ score of 0.895, calculated from multiple evaluation metrics (MUC, B³ and CEAF scores). These results are comparable to the best systems of the challenge. What makes our proposed system distinct is that it also achieves high average F₁ scores for each individual chain type (Test: 0.897, Person: 0.852, Problem: 0.855, Treatment: 0.884). Unlike other works, it obtains good scores for each of the individual metrics rather than being biased towards a particular metric.     

De-identification of unstructured paper-based health records for privacy-preserving secondary use

Whenever personal data is processed, privacy is a serious issue. Especially in the document-centric e-health area, the patients’ privacy must be preserved in order to prevent any negative repercussions for the patient. Clinical research, for example, demands structured health records to carry out efficient clinical trials, whereas legislation (e.g. HIPAA) regulates that only de-identified health records may be used for research. However, unstructured and often paper-based data dominates information technology, especially in the healthcare sector. Existing approaches are geared towards data in English-language documents only and have not been designed to handle the recognition of erroneous personal data which is the result of the OCR-based digitization of paper-based health records.     

Classification of clinically useful sentences in clinical evidence resources

Most patient care questions raised by clinicians can be answered by online clinical knowledge resources. However, important barriers still challenge the use of these resources at the point of care.ObjectiveTo design and assess a method for extracting clinically useful sentences from synthesized online clinical resources that represent the most clinically useful information for directly answering clinicians’ information needs.Materials and methodsWe developed a Kernel-based Bayesian Network classification model based on different domain-specific feature types extracted from sentences in a gold standard composed of 18 UpToDate documents. These features included UMLS concepts and their semantic groups, semantic predications extracted by SemRep, patient population identified by a pattern-based natural language processing (NLP) algorithm, and cue words extracted by a feature selection technique. Algorithm performance was measured in terms of precision, recall, and F-measure.ResultsThe feature-rich approach yielded an F-measure of 74% versus 37% for a feature co-occurrence method (p < 0.001). Excluding predication, population, semantic concept or text-based features reduced the F-measure to 62%, 66%, 58% and 69% respectively (p < 0.01). The classifier applied to Medline sentences reached an F-measure of 73%, which is equivalent to the performance of the classifier on UpToDate sentences (p = 0.62).ConclusionsThe feature-rich approach significantly outperformed general baseline methods. This approach significantly outperformed classifiers based on a single type of feature. Different types of semantic features provided a unique contribution to overall classification performance. The classifier’s model and features used for UpToDate generalized well to Medline abstracts.     

Supervised methods for symptom name recognition in free-text clinical records of traditional Chinese medicine: An empirical study

Clinical records of traditional Chinese medicine (TCM) are documented by TCM doctors during their routine diagnostic work. These records contain abundant knowledge and reflect the clinical experience of TCM doctors. In recent years, with the modernization of TCM clinical practice, these clinical records have begun to be digitized. Data mining (DM) and machine learning (ML) methods provide an opportunity for researchers to discover TCM regularities buried in the large volume of clinical records. There has been some work on this problem. Existing methods have been validated on a limited amount of manually well-structured data. However, the contents of most fields in the clinical records are unstructured. As a result, the previous methods verified on the well-structured data will not work effectively on the free-text clinical records (FCRs), and the FCRs are, consequently, required to be structured in advance. Manually structuring the large volume of TCM FCRs is time-consuming and labor-intensive, but the development of automatic methods for the structuring task is at an early stage. Therefore, in this paper, symptom name recognition (SNR) in the chief complaints, which is one of the important tasks to structure the FCRs of TCM, is carefully studied. The SNR task is reasonably treated as a sequence labeling problem, and several fundamental and practical problems in the SNR task are studied, such as how to adapt a general sequence labeling strategy for the SNR task according to the domain-specific characteristics of the chief complaints and which sequence classifier is more appropriate to solve the SNR task. To answer these questions, a series of elaborate experiments were performed, and the results are explained in detail.     

稳态视觉诱发电位频率识别算法研究进展

稳态视觉诱发电位是脑-机接口系统中常用的控制信号之一。基于稳态视觉诱发电位的脑-机接口具有高信息传输率、训练时间短等优点,已成为脑-机接口研究领域的一个重要分支。本文从无监督学习算法、有监督学习算法和深度学习算法三个方面,归纳了近五年稳态视觉诱发电位频率识别算法研究的主要进展,并展望了值得关注的若干前沿问题和研究方向。     

Development and evaluation of RapTAT: A machine learning system for concept mapping of phrases from medical narratives

Rapid, automated determination of the mapping of free text phrases to pre-defined concepts could assist in the annotation of clinical notes and increase the speed of natural language processing systems. The aim of this study was to design and evaluate a token-order-specific naïve Bayes-based machine learning system (RapTAT) to predict associations between phrases and concepts. Performance was assessed using a reference standard generated from 2860 VA discharge summaries containing 567,520 phrases that had been mapped to 12,056 distinct Systematized Nomenclature of Medicine – Clinical Terms (SNOMED CT) concepts by the MCVS natural language processing system. It was also assessed on the manually annotated, 2010 i2b2 challenge data. Performance was established with regard to precision, recall, and F-measure for each of the concepts within the VA documents using bootstrapping. Within that corpus, concepts identified by MCVS were broadly distributed throughout SNOMED CT, and the token-order-specific language model achieved better performance based on precision, recall, and F-measure (0.95 ± 0.15, 0.96 ± 0.16, and 0.95 ± 0.16, respectively; mean ± SD) than the bag-of-words based, naïve Bayes model (0.64 ± 0.45, 0.61 ± 0.46, and 0.60 ± 0.45, respectively) that has previously been used for concept mapping. Precision, recall, and F-measure on the i2b2 test set were 92.9%, 85.9%, and 89.2% respectively, using the token-order-specific model. RapTAT required just 7.2 ms to map all phrases within a single discharge summary, and mapping rate did not decrease as the number of processed documents increased. The high performance attained by the tool in terms of both accuracy and speed was encouraging, and the mapping rate should be sufficient to support near-real-time, interactive annotation of medical narratives. These results demonstrate the feasibility of rapidly and accurately mapping phrases to a wide range of medical concepts based on a token-order-specific naïve Bayes model and machine learning.     

Word sense disambiguation across two domains: biomedical literature and clinical notes

The aim of this study is to explore the word sense disambiguation (WSD) problem across two biomedical domains—biomedical literature and clinical notes. A supervised machine learning technique was used for the WSD task. One of the challenges addressed is the creation of a suitable clinical corpus with manual sense annotations. This corpus in conjunction with the WSD set from the National Library of Medicine provided the basis for the evaluation of our method across multiple domains and for the comparison of our results to published ones. Noteworthy is that only 20% of the most relevant ambiguous terms within a domain overlap between the two domains, having more senses associated with them in the clinical space than in the biomedical literature space. Experimentation with 28 different feature sets rendered a system achieving an average F-score of 0.82 on the clinical data and 0.86 on the biomedical literature.     

Comparison of automatic summarisation methods for clinical free text notes

ObjectiveA major source of information available in electronic health record (EHR) systems are the clinical free text notes documenting patient care. Managing this information is time-consuming for clinicians. Automatic text summarisation could assist clinicians in obtaining an overview of the free text information in ongoing care episodes, as well as in writing final discharge summaries. We present a study of automated text summarisation of clinical notes. It looks to identify which methods are best suited for this task and whether it is possible to automatically evaluate the quality differences of summaries produced by different methods in an efficient and reliable way.Methods and materialsThe study is based on material consisting of 66,884 care episodes from EHRs of heart patients admitted to a university hospital in Finland between 2005 and 2009. We present novel extractive text summarisation methods for summarising the free text content of care episodes. Most of these methods rely on word space models constructed using distributional semantic modelling. The summarisation effectiveness is evaluated using an experimental automatic evaluation approach incorporating well-known ROUGE measures. We also developed a manual evaluation scheme to perform a meta-evaluation on the ROUGE measures to see if they reflect the opinions of health care professionals.ResultsThe agreement between the human evaluators is good (ICC = 0.74, p < 0.001), demonstrating the stability of the proposed manual evaluation method. Furthermore, the correlation between the manual and automated evaluations are high (> 0.90 Spearman's rho). Three of the presented summarisation methods (‘Composite’, ‘Case-Based’ and ‘Translate’) significantly outperform the other methods for all ROUGE measures (p < 0.05, Wilcoxon signed-rank test and Bonferroni correction).ConclusionThe results indicate the feasibility of the automated summarisation of care episodes. Moreover, the high correlation between manual and automated evaluations suggests that the less labour-intensive automated evaluations can be used as a proxy for human evaluations when developing summarisation methods. This is of significant practical value for summarisation method development, because manual evaluation cannot be afforded for every variation of the summarisation methods. Instead, one can resort to automatic evaluation during the method development process.     

Automated misspelling detection and correction in clinical free-text records

Accurate electronic health records are important for clinical care and research as well as ensuring patient safety. It is crucial for misspelled words to be corrected in order to ensure that medical records are interpreted correctly. This paper describes the development of a spelling correction system for medical text. Our spell checker is based on Shannon’s noisy channel model, and uses an extensive dictionary compiled from many sources. We also use named entity recognition, so that names are not wrongly corrected as misspellings. We apply our spell checker to three different types of free-text data: clinical notes, allergy entries, and medication orders; and evaluate its performance on both misspelling detection and correction. Our spell checker achieves detection performance of up to 94.4% and correction accuracy of up to 88.2%. We show that high-performance spelling correction is possible on a variety of clinical documents.     

DEEPEN: A negation detection system for clinical text incorporating dependency relation into NegEx

In Electronic Health Records (EHRs), much of valuable information regarding patients’ conditions is embedded in free text format. Natural language processing (NLP) techniques have been developed to extract clinical information from free text. One challenge faced in clinical NLP is that the meaning of clinical entities is heavily affected by modifiers such as negation. A negation detection algorithm, NegEx, applies a simplistic approach that has been shown to be powerful in clinical NLP. However, due to the failure to consider the contextual relationship between words within a sentence, NegEx fails to correctly capture the negation status of concepts in complex sentences. Incorrect negation assignment could cause inaccurate diagnosis of patients’ condition or contaminated study cohorts. We developed a negation algorithm called DEEPEN to decrease NegEx’s false positives by taking into account the dependency relationship between negation words and concepts within a sentence using Stanford dependency parser. The system was developed and tested using EHR data from Indiana University (IU) and it was further evaluated on Mayo Clinic dataset to assess its generalizability. The evaluation results demonstrate DEEPEN, which incorporates dependency parsing into NegEx, can reduce the number of incorrect negation assignment for patients with positive findings, and therefore improve the identification of patients with the target clinical findings in EHRs.     

电子病历中命名实体的智能识别

电子病历中命名实体的识别对于构建和挖掘大型临床数据库以服务于临床决策具有重要意义,而我国目前对此的研究相对较少。在比较现有的实体识别方法和模型后,采用条件随机场模型(CRF)机器学习的方法,对疾病、临床症状、手术操作3类中文病历中常见的命名实体进行智能识别。首先,通过分析电子病历的数据特征,选择以语言符号、词性、构词特征、词边界、上下文为特征集。然后,基于随机抽取的来自临床医院多个科室的电子病历数据,构建小规模语料库并进行标注。最后,利用条件随机场算法执行工具CRF++进行3次对照实验。通过逐步分析特征集中的多种特征对CRF自动识别的影响,提出在中文病历环境下CRF特征选择和模板设计的一些基本规则。在对照实验中,本方法取得了良好效果,3类实体的最佳F值分别达到了92.67%、93.76%和95.06%。     

Using machine learning methods for predicting inhospital mortality in patients undergoing open repair of abdominal aortic aneurysm

An abdominal aortic aneurysm is an abnormal dilatation of the aortic vessel at abdominal level. This disease presents high rate of mortality and complications causing a decrease in the quality of life and increasing the cost of treatment. To estimate the mortality risk of patients undergoing surgery is complex due to the variables associated. The use of clinical decision support systems based on machine learning could help medical staff to improve the results of surgery and get a better understanding of the disease. In this work, the authors present a predictive system of inhospital mortality in patients who were undergoing to open repair of abdominal aortic aneurysm. Different methods as multilayer perceptron, radial basis function and Bayesian networks are used. Results are measured in terms of accuracy, sensitivity and specificity of the classifiers, achieving an accuracy higher than 95%. The developing of a system based on the algorithms tested can be useful for medical staff in order to make a better planning of care and reducing undesirable surgery results and the cost of the post-surgical treatments.