scholarly journals The Comparative Experimental Study of Multilabel Classification for Diagnosis Assistant Based on Chinese Obstetric EMRs

2018 ◽  
Vol 2018 ◽  
pp. 1-9 ◽  
Author(s):  
Kunli Zhang ◽  
Hongchao Ma ◽  
Yueshu Zhao ◽  
Hongying Zan ◽  
Lei Zhuang
Keyword(s):  
Electronic Medical Records ◽  
Medical Records ◽  
Latent Dirichlet Allocation ◽  
Nearest Neighbor ◽  
Fertility Level ◽  
K Nearest Neighbor ◽  
Multilabel Classification ◽  
Multilabel Learning ◽  
Chief Complaints ◽  
Physical Examinations

Obstetric electronic medical records (EMRs) contain massive amounts of medical data and health information. The information extraction and diagnosis assistants of obstetric EMRs are of great significance in improving the fertility level of the population. The admitting diagnosis in the first course record of the EMR is reasoned from various sources, such as chief complaints, auxiliary examinations, and physical examinations. This paper treats the diagnosis assistant as a multilabel classification task based on the analyses of obstetric EMRs. The latent Dirichlet allocation (LDA) topic and the word vector are used as features and the four multilabel classification methods, BP-MLL (backpropagation multilabel learning), RAkEL (RAndom k labELsets), MLkNN (multilabel k-nearest neighbor), and CC (chain classifier), are utilized to build the diagnosis assistant models. Experimental results conducted on real cases show that the BP-MLL achieves the best performance with an average precision up to 0.7413 ± 0.0100 when the number of label sets and the word dimensions are 71 and 100, respectively. The result of the diagnosis assistant can be introduced as a supplementary learning method for medical students. Additionally, the method can be used not only for obstetric EMRs but also for other medical records.

scholarly journals Exploiting Language Models to Classify Events from Twitter

2015 ◽  
Vol 2015 ◽  
pp. 1-11 ◽  
Author(s):  
Duc-Thuan Vo ◽  
Vo Thuan Hai ◽  
Cheol-Young Ock
Keyword(s):  
Latent Dirichlet Allocation ◽  
Nearest Neighbor ◽  
Language Models ◽  
K Nearest Neighbor ◽  
Text Corpora ◽  
Common Term ◽  
Selectional Preferences ◽  
Linguistic Relations ◽  
Relationship Of ◽  
Learning Language

Classifying events is challenging in Twitter because tweets texts have a large amount of temporal data with a lot of noise and various kinds of topics. In this paper, we propose a method to classify events from Twitter. We firstly find the distinguishing terms between tweets in events and measure their similarities with learning language models such as ConceptNet and a latent Dirichlet allocation method for selectional preferences (LDA-SP), which have been widely studied based on large text corpora within computational linguistic relations. The relationship of term words in tweets will be discovered by checking them under each model. We then proposed a method to compute the similarity between tweets based on tweets’ features including common term words and relationships among their distinguishing term words. It will be explicit and convenient for applying to k-nearest neighbor techniques for classification. We carefully applied experiments on the Edinburgh Twitter Corpus to show that our method achieves competitive results for classifying events.

scholarly journals Validation of the Michigan’s Public Health Syndromic System Using Electronic Medical Records

2015 ◽  
Vol 7 (1) ◽  
Author(s):  
Sandhya Swarnavel ◽  
Jim Collins ◽  
Corrine Miller
Keyword(s):  
Public Health ◽  
Electronic Medical Records ◽  
Medical Records ◽  
Situational Awareness ◽  
Communicable Disease ◽  
Tertiary Hospital ◽  
Chief Complaint ◽  
Automated Classification ◽  
Chief Complaints

Michigan has been collecting chief complaint data from emergency departments statewide to support situational awareness activities related to communicable disease since 2004. We validated the syndromic system by comparing the chief complaint data to the electronic medical records (EMR) of a tertiary hospital in southeast Michigan to better understand the utility of the system for non-communicable disease situations.Findings of this study will help determine the accuracy of the automated classification of data based on chief complaints. This study can add confidence in planning for public health preparedness activities and situational awareness.

scholarly journals Harvesting Online Reviews to Identify the Competitor Set in a Service Business: Evidence From the Hotel Industry

2020 ◽  
pp. 109467052097514
Author(s):  
Fei Ye ◽  
Qian Xia ◽  
Minhao Zhang ◽  
Yuanzhu Zhan ◽  
Yina Li
Keyword(s):  
Latent Dirichlet Allocation ◽  
Nearest Neighbor ◽  
Service Industry ◽  
Analytical Framework ◽  
Online Reviews ◽  
Service Industries ◽  
K Nearest Neighbor ◽  
Allocation Model ◽  
Customer Reviews ◽  
Latent Dirichlet Allocation Model

In today’s global service industry, online reviews posted by consumers offer critical information that influences subsequent consumers’ purchasing decisions and firms’ operation strategies. However, little research has been done on how the same information can be used to identify key competitors and improve services to increase competitiveness. In this article, we propose an analytical framework based on an improved k-nearest neighbor model and a latent Dirichlet allocation model for service managers to harvest online reviews to identify their key competitors and to evaluate the strengths and weaknesses of their businesses. With a sample comprising over 8 million customer reviews of 6,409 hotels in 50 Chinese cities from Ctrip.com , we validate the effectiveness of the proposed approach in the analysis of a hotel’s service competitiveness and its key competitors. The findings indicate that the importance of particular attributes of a hotel varies in different segments according to hotel star ratings. This study extends the literature by bridging online reviews and competitor identification for service industries. It also contributes to practice by offering a systematic and effective way for managers to identify their key competitors, monitor market preferences, ensure service quality, and formulate effective marketing strategies.

scholarly journals Essential Data Elements for Electronic Cardiovascular Medical Record Systems in Iran

2021 ◽  
Vol 10 (1) ◽  
pp. 54
Author(s):  
Abbas Sheikhtaheri ◽  
Farid Khorami ◽  
Hedyeh Mohammadzadeh
Keyword(s):  
Medical Record ◽  
Electronic Medical Records ◽  
Medical Records ◽  
Clinical Signs ◽  
Minimum Data Set ◽  
Nursing Interventions ◽  
Data Set ◽  
Minimum Data ◽  
Physical Examinations ◽  
Data Elements

Introduction: Electronic medical records play an important role in the management of patients. In order to develop cardiovascular electronic medical record systems, determining minimum data set is necessary. This study aimed to determine the essential data elements for electronic cardiovascular medical record systems.Methods: Medical records of patients with cardiovascular diseases and also the literature were reviewed to develop a questionnaire regarding the data elements.  87 cardiovascular specialists and residents as well as 50 nurses working in cardiovascular departments of hospitals affiliated with Iran University of Medical Sciences participated in the study. The data elements with at least 75% of agreement were considered essential for electronic medical records. Data were analyzed using descriptive statistics in SPSS software.Results: The essential  data elements were classified in 29 classes including admission, death, patients’ main complaints, clinical signs, observations, medications, cardiac surgery, risk factors, laboratory and pathology results, consultation, resuscitation, anesthetic, electrocardiography, blood transfusion or blood products, rehabilitation measures, angiography/venography, exercise testing, endoscopy/colonoscopy, medical imaging, echocardiography, nursing interventions, allergies and side effects, therapeutic implantations, cardiac examinations, physical examinations, angina, referrals, social backgrounds and history., Totally, out of 276 data elements, 245 elements were identified as the essential data elements for electronic cardiovascular medical record systems.Conclusion: In this study, essential data elements were defined for electronic cardiovascular medical records. Identifying cardiovascular minimum data set will be an effective step towards integrating and improving the management of these patients' information.

scholarly journals Multi-Self-Attention for Aspect Category Detection and Biomedical Multilabel Text Classification with BERT

2021 ◽  
Vol 2021 ◽  
pp. 1-6
Author(s):  
Xuelei Zhang ◽  
Xinyu Song ◽  
Ao Feng ◽  
Zhengjie Gao
Keyword(s):  
Natural Language Processing ◽  
Natural Language ◽  
Electronic Medical Records ◽  
Language Processing ◽  
Medical Records ◽  
Multilabel Classification ◽  
Attention Model ◽  
Natural Choice ◽  
Challenging Tasks ◽  
Word Attention

Multilabel classification is one of the most challenging tasks in natural language processing, posing greater technical difficulties than single-label classification. At the same time, multilabel classification has more natural applications. For individual labels, the whole piece of text has different focuses or component distributions, which require full use of local information of the sentence. As a widely adopted mechanism in natural language processing, attention becomes a natural choice for the issue. This paper proposes a multilayer self-attention model to deal with aspect category and word attention at different granularities. Combined with the BERT pretraining model, it achieves competitive performance in aspect category detection and electronic medical records’ classification.

scholarly journals Study on the semi-supervised learning-based patient similarity from heterogeneous electronic medical records

2021 ◽  
Vol 21 (S2) ◽  
Author(s):  
Ni Wang ◽  
Yanqun Huang ◽  
Honglei Liu ◽  
Zhiqiang Zhang ◽  
Lan Wei ◽  
...  
Keyword(s):  
Supervised Learning ◽  
Electronic Medical Records ◽  
Medical Records ◽  
Euclidean Distance ◽  
Learning Algorithm ◽  
Characteristic Curve ◽  
Similarity Measurement ◽  
K Nearest Neighbor ◽  
Training Samples ◽  
Similarity Scores

Abstract Background A new learning-based patient similarity measurement was proposed to measure patients’ similarity for heterogeneous electronic medical records (EMRs) data. Methods We first calculated feature-level similarities according to the features’ attributes. A domain expert provided patient similarity scores of 30 randomly selected patients. These similarity scores and feature-level similarities for 30 patients comprised the labeled sample set, which was used for the semi-supervised learning algorithm to learn the patient-level similarities for all patients. Then we used the k-nearest neighbor (kNN) classifier to predict four liver conditions. The predictive performances were compared in four different situations. We also compared the performances between personalized kNN models and other machine learning models. We assessed the predictive performances by the area under the receiver operating characteristic curve (AUC), F1-score, and cross-entropy (CE) loss. Results As the size of the random training samples increased, the kNN models using the learned patient similarity to select near neighbors consistently outperformed those using the Euclidean distance to select near neighbors (all P values < 0.001). The kNN models using the learned patient similarity to identify the top k nearest neighbors from the random training samples also had a higher best-performance (AUC: 0.95 vs. 0.89, F1-score: 0.84 vs. 0.67, and CE loss: 1.22 vs. 1.82) than those using the Euclidean distance. As the size of the similar training samples increased, which composed the most similar samples determined by the learned patient similarity, the performance of kNN models using the simple Euclidean distance to select the near neighbors degraded gradually. When exchanging the role of the Euclidean distance, and the learned patient similarity in selecting the near neighbors and similar training samples, the performance of the kNN models gradually increased. These two kinds of kNN models had the same best-performance of AUC 0.95, F1-score 0.84, and CE loss 1.22. Among the four reference models, the highest AUC and F1-score were 0.94 and 0.80, separately, which were both lower than those for the simple and similarity-based kNN models. Conclusions This learning-based method opened an opportunity for similarity measurement based on heterogeneous EMR data and supported the secondary use of EMR data.

scholarly journals Zenker Diverticulum: Does Size Correlate With Preoperative Symptoms?

2021 ◽  
Author(s):  
Jhon F. Martinez-Paredes ◽  
Razan Alfakir ◽  
Jan L. Kasperbauer ◽  
Amy Rutt
Keyword(s):  
Quality Of Life ◽  
Electronic Medical Records ◽  
Medical Records ◽  
Gastrointestinal Symptoms ◽  
Zenker Diverticulum ◽  
Chief Complaints ◽  
Upper Gastrointestinal Symptoms ◽  
Retrospective Study Design ◽  
Upper Gastrointestinal

Abstract Introduction Zenker diverticulum (ZD) usually affects adults after the 7th decade of life. Treatment for ZD is indicated for all symptomatic patients, but some patients prefer to defer surgical treatment until symptoms get worse and decrease their quality of life. Objective To evaluate the association of the preoperative symptoms in ZD patients with the size of the ZD. Methods A retrospective study design. Electronic medical records were used to identify patients diagnosed with ZD and treated over 11 years. Data collection included the chief complaints and symptoms, medical history, and findings on radiologic swallow evaluations of the patients. The diverticulum size was stratified into 3 groups: small (< 1 cm), moderate (1–3 cm), and large (> 3 cm). Results A total of 165 patients were enrolled and stratified by diverticulum size (48 small, 67 medium, and 50 large). Dysphagia, cough, and regurgitation were the most prevalent symptoms. Dysphonia was more frequent among patients with a small pouch. Logistic regression analysis showed that dysphagia and choking were associated with large and medium diverticulum size (p < 0.05). Additionally, dysphonia was significantly associated with the presence of a small-sized ZD (p < 0.04). Conclusion Upper gastrointestinal symptoms such as dysphagia and choking may be associated with a ZD > 1 cm and should always be evaluated. Additionally, the presence of dysphonia was found to be correlated with a ZD < 1 cm, suggesting that a prompt and appropriate fluoroscopic evaluation must be considered in those patients in whom no other clear cause of dysphonia is evident.

A brief intervention to improve rates and quality of physical examinations for admissions to acute adult psychiatry units

2019 ◽  
Vol 27 (6) ◽  
pp. 641-644
Author(s):  
Usama Munir ◽  
Adnan Younus ◽  
Vlasios Brakoulias
Keyword(s):  
Physical Examination ◽  
Electronic Medical Records ◽  
Brief Intervention ◽  
Retrospective Review ◽  
Medical Records ◽  
Breath Sounds ◽  
Adult Psychiatry ◽  
Physical Examinations ◽  
The Mean

Objective: To determine the frequency and quality of physical examinations within 24 h of admission to an acute adult psychiatry unit, and whether a brief intervention involving feedback to clinicians could lead to improvement. Method: Retrospective review of the electronic medical records followed by four brief feedback sessions and email correspondence, followed by a further review of the medical records 1 month later. Results: The proportion of patients receiving a physical examination increased from 36/71 (50.7%) in the initial audit to 41/64 (64.1%) in the re-audit. The mean score of the quality of physical examinations improved from 7.5 to 9.3 (out of 15). The greatest improvement on re-audit occurred in the documentation of additional cardiac sounds (33.9% increase), additional breath sounds (17.7% increase), breath sounds (17.1% increase), cardiac sounds (14.2% increase) and bowel sounds (12.5% increase). Conclusion: This audit supports the use of brief peer-led feedback to improve the rates and quality of physical examinations.

Sign in / Sign up

Export Citation Format

Share Document