scholarly journals Distinct psychopathology profiles in patients with epileptic seizures compared to non-epileptic psychogenic seizures

2019 ◽  
Author(s):  
Albert D Wang ◽  
Michelle Leong ◽  
Benjamin Johnstone ◽  
Genevieve Rayner ◽  
Tomas Kalincik ◽  
...  
Keyword(s):  
Classification Accuracy ◽  
Linear Models ◽  
Learning Algorithm ◽  
Epileptic Seizures ◽  
Roc Curves ◽  
Screening Tools ◽  
General Anxiety ◽  
Phobic Anxiety ◽  
Domain Scores ◽  
R Domain

AbstractObjectiveSimilarities in clinical presentations between epileptic seizures (ES) and psychogenic non-epileptic seizures (PNES) produces a risk of misdiagnosis. Video-EEG monitoring (VEM) is the diagnostic gold standard, but involves significant cost and time commitment, suggesting a need for efficient screening tools.Methods628 patients were recruited from an inpatient VEM unit; 293 patients with ES, 158 with PNES, 31 both ES and PNES, and 146 non-diagnostic. Patients completed the SCL-90-R, a standardised 90-item psychopathology instrument. Bayesian linear models were computed to investigate whether SCL-90-R domain scores or the overall psychopathology factor p differed between groups. Receiver operating characteristic (ROC) curves were computed to investigate the PNES classification accuracy of each domain score and p. A machine learning algorithm was also used to determine which subset of SCL-90-R items produced the greatest classification accuracy.ResultsEvidence was found for elevated scores in PNES compared to ES groups in the symptom domains of anxiety (b = 0.47, 95%HDI = [0.10, 0.80]), phobic anxiety (b = 1.32, 95%HDI = [0.98, 1.69]), somatisation (b = 0.84, 95%HDI = [0.49, 1.20]), and the general psychopathology factor p (b = 1.35, 95%HDI = [0.86, 1.82]). Of the SCL-90-R domain scores, somatisation produced the highest classification accuracy (AUC = 0.74, 95%CI = [0.69, 0.79]). The genetic algorithm produced a 6-item subset from the SCL-90-R, which produced comparable classification accuracy to the somatisation scores (AUC = 0.73, 95%CI = [0.64, 0.82]).SignificanceCompared to patients with ES, patients with PNES report greater symptoms of somatisation, general anxiety, and phobic anxiety against a background of generally elevated psychopathology. While self-reported psychopathology scores are not accurate enough for diagnosis in isolation, elevated psychopathology in these domains should raise the suspicion of PNES in clinical settings.

scholarly journals The Child Behavior Checklist as a Screening Instrument for PTSD in Refugee Children

Children ◽  
2021 ◽  
Vol 8 (6) ◽  
pp. 521
Author(s):  
Ina Nehring ◽  
Heribert Sattel ◽  
Maesa Al-Hallak ◽  
Martin Sack ◽  
Peter Henningsen ◽  
...  
Keyword(s):  
Child Behavior ◽  
Child Behavior Checklist ◽  
High Sensitivity ◽  
Roc Curves ◽  
Structured Interview ◽  
Screening Instrument ◽  
Screening Tools ◽  
Refugee Children ◽  
Post Traumatic Stress ◽  
Sensitivity Specificity

Thousands of refugees who have entered Europe experienced threatening conditions, potentially leading to post traumatic stress disorder (PTSD), which has to be detected and treated early to avoid chronic manifestation, especially in children. We aimed to evaluate and test suitable screening tools to detect PTSD in children. Syrian refugee children aged 4–14 years were examined using the PTSD-semi-structured interview, the Kinder-DIPS, and the Child Behavior Checklist (CBCL). The latter was evaluated as a potential screening tool for PTSD using (i) the CBCL-PTSD subscale and (ii) an alternative subscale consisting of a psychometrically guided selection of items with an appropriate correlation to PTSD and a sufficient prevalence (presence in more than 20% of the cases with PTSD). For both tools we calculated sensitivity, specificity, and a receiver operating characteristic (ROC) curve. Depending on the sum score of the items, the 20-item CBCL-PTSD subscale as used in previous studies yielded a maximal sensitivity of 85% and specificity of 76%. The psychometrically guided item selection resulted in a sensitivity of 85% and a specificity of 83%. The areas under the ROC curves were the same for both tools (0.9). Both subscales may be suitable as screening instrument for PTSD in refugee children, as they reveal a high sensitivity and specificity.

scholarly journals Real-Time AI-Based Informational Decision-Making Support System Utilizing Dynamic Text Sources

2021 ◽  
Vol 11 (13) ◽  
pp. 6237
Author(s):  
Azharul Islam ◽  
KyungHi Chang
Keyword(s):  
Machine Learning ◽  
Decision Making ◽  
Random Forest ◽  
Support System ◽  
Classification Accuracy ◽  
Short Term Memory ◽  
Learning Algorithm ◽  
Unstructured Data ◽  
Stochastic Gradient Descent ◽  
Decision Making Support

Unstructured data from the internet constitute large sources of information, which need to be formatted in a user-friendly way. This research develops a model that classifies unstructured data from data mining into labeled data, and builds an informational and decision-making support system (DMSS). We often have assortments of information collected by mining data from various sources, where the key challenge is to extract valuable information. We observe substantial classification accuracy enhancement for our datasets with both machine learning and deep learning algorithms. The highest classification accuracy (99% in training, 96% in testing) was achieved from a Covid corpus which is processed by using a long short-term memory (LSTM). Furthermore, we conducted tests on large datasets relevant to the Disaster corpus, with an LSTM classification accuracy of 98%. In addition, random forest (RF), a machine learning algorithm, provides a reasonable 84% accuracy. This research’s main objective is to increase the application’s robustness by integrating intelligence into the developed DMSS, which provides insight into the user’s intent, despite dealing with a noisy dataset. Our designed model selects the random forest and stochastic gradient descent (SGD) algorithms’ F1 score, where the RF method outperforms by improving accuracy by 2% (to 83% from 81%) compared with a conventional method.

scholarly journals Specific mental disorder screening compilation may detect general mental disorders

2018 ◽  
Vol 13 (40) ◽  
pp. 1-13
Author(s):  
Lívia Maria Bolsoni ◽  
Leonardo Moscovici ◽  
João Mazzoncini De Azevedo Marques ◽  
Antonio Waldo Zuardi
Keyword(s):  
General Population ◽  
Mental Disorders ◽  
Emotional Disorders ◽  
Intraclass Correlation ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Screening Tools ◽  
Patient Health ◽  
Two Phases ◽  
Specific Disorders

Objective: To evaluate whether a short compilation of screening tools for specific disorders could identify Mental or Emotional Disorders (MEDs) in the general population. Methods: We selected validated screening tools for the most prevalent MEDs. In order to be selected, these tools should maintain the psychometric properties of the complete instrument with a reduced number of items. These instruments were: Patient Health Questionnaire-2 (PHQ-2), Generalized Anxiety Disorder Scale-2 (GAD-2), item 3 of the Alcohol Use Disorders Identification Test (AUDIT), and three items on the Adolescent Psychotic-Like Symptom Screener (APSS-3). We called this compilation of screening tools Mini Screening for Mental Disorders (Mini-SMD). The study was divided in two phases. Firstly, 545 subjects were interviewed with the Mini-SMD and COOP/WONCA-Feelings at their residences. Subsequently, subjects who had agreed to participate (230) were reinterviewed with Mini-SMD, COOP/WONCA-Feelings and MINI interview. Test-retest reliability was calculated by Intraclass Correlation Coefficient (ICC). Receiver operating characteristic (ROC) curves were generated for the analysis of discriminative validity. Concurrent validity was calculated by analyzing the correlation between Mini-SMD and COOP/WONCA-Feelings. Results: The joint administration of screening tools for specific disorders showed sensitivities that ranged from 0.76 to 0.88 and specificities from 0.67 to 0.85. The ICC value for the total score of Mini-SMD was 0.78. The area under the curve was 0.84, with a sensitivity of 0.74 and specificity of 0.76 (for a cutoff ≥ 4). Conclusion: This study showed that a short compilation of screening tools for specific disorders can detect MEDs in general population.

Deep Learning Based Adaptive Recurrent Neural Network for Detection of Myocardial Infarction

2021 ◽  
Vol 11 (12) ◽  
pp. 3044-3053
Author(s):  
Rakesh Kumar Mahendran ◽  
V. Prabhu ◽  
V. Parthasarathy ◽  
A. Mary Judith
Keyword(s):  
Neural Network ◽  
Myocardial Infarction ◽  
Deep Learning ◽  
Recurrent Neural Network ◽  
Classification Accuracy ◽  
Learning Algorithm ◽  
Training Data ◽  
Discrete Wavelet ◽  
Ecg Signal ◽  
Detection Techniques

Myocardial infarction (MI) may precipitate severe health damage and lead to irreversible death of the heart muscle, the result of prolonged lack of oxygen if it is not treated in a timely manner. Lack of accurate and early detection techniques for this heart disease has reduced the efficiency of MI diagnosis. In this paper, the design, and implementation of an efficient deep learning algorithm called Adaptive Recurrent neural network (ARNN) is proposed for the MI detection. The main objective of the proposed work is the accurate identification of MI disease using ECG signals. ECG signal denoising has been performed using the Multi-Notch filter, which removes the specified noise frequency range. Discrete wavelet transform (DWT) is utilized for performing the feature extraction that decomposes the ECG signal into varied scales with waveletfiltering bank. After the extraction of specific QRS features, classification of the defected and normal ECG arrhythmic beat has been performed using the deep learning-based ARNN classifier. The MIT-BIH database has been used for testing and training data. The performance of the proposed algorithm is evaluated based on classification accuracy. Results that are attained include the classification accuracy of about 99.21%, 99% of sensitivity and 99.4% of specificity with PPV and NPV of about 99.4 and 99.01 values indicate the enhanced performance of our proposed work compared with the conventional LSTM-CAE and LSTM-CNN techniques.

Hybridization of Machine Learning Algorithm in Intrusion Detection System

2020 ◽  
pp. 150-175
Author(s):  
Amudha P. ◽  
Sivakumari S.
Keyword(s):  
Machine Learning ◽  
Intrusion Detection ◽  
Classification Accuracy ◽  
Intrusion Detection System ◽  
Learning Algorithm ◽  
Detection System ◽  
Principal Component ◽  
Machine Learning Algorithms ◽  
Feature Selection Technique ◽  
Efficient Manner

In recent years, the field of machine learning grows very fast both on the development of techniques and its application in intrusion detection. The computational complexity of the machine learning algorithms increases rapidly as the number of features in the datasets increases. By choosing the significant features, the number of features in the dataset can be reduced, which is critical to progress the classification accuracy and speed of algorithms. Also, achieving high accuracy and detection rate and lowering false alarm rates are the major challenges in designing an intrusion detection system. The major motivation of this work is to address these issues by hybridizing machine learning and swarm intelligence algorithms for enhancing the performance of intrusion detection system. It also emphasizes applying principal component analysis as feature selection technique on intrusion detection dataset for identifying the most suitable feature subsets which may provide high-quality results in a fast and efficient manner.

Determination and Interpretation of the Optimal Operating Point for ROC Curves Derived Through Generalized Linear Models

2003 ◽  
Vol 2 (4) ◽  
pp. 219-242 ◽  
Author(s):  
Robert J. Gallop ◽  
Paul Crits-Christoph ◽  
Larry R. Muenz ◽  
Xin M. Tu
Keyword(s):  
Generalized Linear Models ◽  
Linear Models ◽  
Roc Curves ◽  
Operating Point ◽  
Optimal Operating

Patient-reported health and 1-year mortality in patients with ischemic heart disease – findings from the Denheart study

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
T Rasmussen ◽  
B Borregaard ◽  
P Palm ◽  
R Mols ◽  
A.V Christensen ◽  
...  
Keyword(s):  
Heart Disease ◽  
Ischemic Heart Disease ◽  
Roc Curves ◽  
Screening Tools ◽  
Ischemic Heart ◽  
Time To Death ◽  
Patient Reported ◽  
Returning Home ◽  
One Year ◽  
Pro Instruments

Abstract Background Though survival has improved markedly in ischemic heart disease (IHD), it remains a leading cause of death worldwide. Screening tools to identify patients at risk are ever in demand. Large-scale studies exploring the association between patients' self-reported mental and physical health and mortality are lacking. Purpose (i) to describe patient-reported outcomes (PROs) at discharge in IHD patients deceased and alive at one year, (ii) to investigate the discriminant predictive performance of PRO instruments on mortality, (iii) to investigate differences in time to death among survey responders/non-responders and among three diagnostic sub-groups (chronic ischemic heart disease/stable angina, non-STEMI/unstable angina and STEMI), and (iv) to investigate predictors of one-year mortality among sociodemographic, clinical and self-reported factors. Methods Data from the national DenHeart survey with register-data linkage was used. A total of 14,115 adults with IHD were discharged during one year. Eligible (n=13,476) were invited to complete a questionnaire and 7,167 (53%) responded. Questionnaires included the Health survey short form 12-items (SF-12), Hospital Anxiety and Depression Scale (HADS), EuroQoL-5-dimensions (EQ-5D), HeartQoL, Edmonton Symptom Assessment Scale (ESAS) and ancillary questions. Clinical and demographic characteristics were obtained from registries as were data on one-year mortality. Comparative analyses investigated differences in PROs, and discriminant PRO-performance was explored by Receiver Operating Characteristics (ROC) curves. Kaplan-Meier survival analysis explored differences in time to death across sub-groups. Predictors of mortality were explored using multifactorially adjusted cox regression analyses with time to death as underlying timescale. Results Highly significant and clinically important differences in PROs were found between those alive and those deceased at one year. The best discriminant performance was observed for the physical component scale of the SF-12 (Area Under the Curve (AUC) 0.706) (Figure 1). One-year mortality among responders and non-responders was 2% and 7%, respectively. Significant differences in time to death was observed between responders and non-responders (p<0.001) and among diagnostic subgroups (p<0.001). Strongest predictors of one-year mortality included STEMI (hazard ratio (HR) 2.9 95% confidence interval (CI) 2.3–3.7), Tu comorbidity index score 3+ (HR 3.6, 95% CI 2.7–4.8) and patient-reported feeling unsafe about returning home from hospital (HR 2.07, 95% CI 1.2–3.61). Conclusions One-year post-discharge mortality was expectedly low, however notably higher in certain subgroups. Though clinical predictors may be difficult to modify, factors such as feeling unsafe about returning home should be addressed at discharge. PRO-performance estimates may guide clinicians and researchers in choosing appropriate predictive patient-reported outcome tools. Figure 1. PRO instruments ROC curves Funding Acknowledgement Type of funding source: None

Colored spun fabric texture representation and application by combining spatial features with frequency features

2020 ◽  
pp. 004051752096140
Author(s):  
Li Yuan ◽  
Xue Gong ◽  
Junping Liu ◽  
Yali Yang ◽  
Muli Liu
Keyword(s):  
Classification Accuracy ◽  
Local Binary Pattern ◽  
Learning Algorithm ◽  
Woven Fabric ◽  
Local Phase ◽  
Texture Representation ◽  
Local Phase Quantization ◽  
Fuzzy Transformation ◽  
Representation Model ◽  
Phase Quantization

Colored spun fabrics are difficult to accurately characterize with a local binary pattern due to texture anisotropy caused by the uneven distribution of dyed fibers. In this paper, we present a texture representation model based on spatial and frequency characteristics. The proposed model takes advantage of the local binary pattern and local phase quantization to extract the texture of woven fabric. Then, the two features are connected in series, and the features of dimension reduction by principal component analysis are used to represent the texture of the fabric image. Finally, the hierarchical hybrid classifier is applied to classify the fabric structure. The experimental results show that the local phase quantization feature is robust to the fuzzy transformation and the texture representation model has a stronger ability of texture description than the single local binary pattern feature, with the average classification accuracy of 97.59% on 336 samples. In addition, compared with the deep learning algorithm, the texture representation algorithm can ensure a high classification accuracy.

Screening for Iron Deficiency: Can We Continue To Avoid Ferritin?.

Blood ◽  
2006 ◽  
Vol 108 (11) ◽  
pp. 3351-3351
Author(s):  
Maria E. Montoya ◽  
Peter R. Van Delden ◽  
M. Tarek Elghetany ◽  
J. David Bessman
Keyword(s):  
Iron Deficiency ◽  
Serum Ferritin ◽  
Blood Count ◽  
Roc Curves ◽  
Screening Tools ◽  
Distribution Width ◽  
Males And Females ◽  
Relative Utility ◽  
Multiple Receiver

Abstract Detection of iron deficiency remains poorly understood and costly due to inappropriate screening. Low ferritin is a definitive diagnosis of iron deficiency, but screening with ferritin is not allowed. Therefore surrogates in the blood count have been used to justify obtaining the serum ferritin. The purpose of this research was to analyze the role of Hemoglobin (Hgb), Mean Corpuscular Volume (MCV), and RBC Distribution Width (RDW) as surrogates in screening for iron deficiency. All 2,563 patients with serum ferritin levels gathered over 12 months were reviewed. The relative utility of Hgb, MCV, and RDW in screening for low ferritin levels was shown through multiple Receiver Operator Characteristic (ROC) curves. 264 patients had a ferritin less than 10 ng/ml and 210 between 11 to 20 ng/ml. Results indicate that when viewed independently MCV correlates most closely to low ferritin as seen in Figure 1. RDW and Hgb in both males and females demonstrate a weaker association though remains of value. Table1 lists the values at which the three screening tools were 95% and 100% sensitive for detecting ferritin levels of 10 ng/ml and below. In contrast the data indicate that for ferritin levels from 11 to 20 ng/ml all three screening variables have poor sensitivity and specificity. This is demonstrated clearly in Figure 2. The data suggest that the most severe iron deficiency (ferritin under 10 ng/ml) can be well predicted by abnormalities in the blood count; however less severe iron deficiency (ferritin 10 to 20 ng/ml) cannot be anticipated from the blood count. The blood count does not appear to be a practical alternative to ferritin for screening for iron deficiency. Table 1: Sreening Variable Sensitivities* 100% Sensitivity 95% Sensitivity *values for ferritin less than 11 ng/ml MCV >98.2 >90.0 RDW <12.2 <13.1 Hgb Males >15.0 >13.7 Hgb Females >14.2 >12.6 Figure 1 Figure 1. Figure 2 Figure 2.

Application of artificial neural network to loan recovery prediction

2016 ◽  
Vol 9 (2) ◽  
pp. 222-238 ◽  
Author(s):  
Amos Olaolu Adewusi ◽  
Tunbosun Biodun Oyedokun ◽  
Mustapha Oyewole Bello
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Interest Rates ◽  
Classification Accuracy ◽  
Learning Algorithm ◽  
Type I ◽  
Ann Model ◽  
Content Type ◽  
Credit Evaluation ◽  
Artificial Neural

Purpose This study assesses the classification accuracy of an artificial neural network (ANN) model. It examines the application of loan recovery probability rather than odds of default as the case with traditional credit evaluation models. Design/methodology/approach Data on 2,300 loans granted over the period 2001-2012 was obtained from the databases of Nigerian commercial banks and primary mortgage institutions. A multilayer feed-forward ANN model with back-propagation learning algorithm was developed having classified the sample into training (38 per cent), testing (41 per cent) and validation (21 per cent) sub-samples. Findings The model exhibits a high overall percentage classification accuracy of 92.6 per cent. It also achieves relatively low misclassification Type I and Type II errors at 6.5 per cent and 8.2 per cent, respectively. Macroeconomic variables such as gross domestic product, inflation and interest rates have the strongest influence on the ANN model classification power. The result of the analysis shows that adopting odds of recovery in ANN classification models can lead to improved loan evaluation. Originality/value The paper is distinct from extant studies in that it presents a new dimension to loan evaluation in Nigerian lending market. To the best knowledge of the authors, the paper is among the first to explore probability of loan recovery as the basis for credit evaluation in the country.

Sign in / Sign up

Export Citation Format

Share Document