Explainable Machine Learning Model for Glaucoma Diagnosis and Its Interpretation

The aim is to develop a machine learning prediction model for the diagnosis of glaucoma and an explanation system for a specific prediction. Clinical data of the patients based on a visual field test, a retinal nerve fiber layer optical coherence tomography (RNFL OCT) test, a general examination including an intraocular pressure (IOP) measurement, and fundus photography were provided for the feature selection process. Five selected features (variables) were used to develop a machine learning prediction model. The support vector machine, C5.0, random forest, and XGboost algorithms were tested for the prediction model. The performance of the prediction models was tested with 10-fold cross-validation. Statistical charts, such as gauge, radar, and Shapley Additive Explanations (SHAP), were used to explain the prediction case. All four models achieved similarly high diagnostic performance, with accuracy values ranging from 0.903 to 0.947. The XGboost model is the best model with an accuracy of 0.947, sensitivity of 0.941, specificity of 0.950, and AUC of 0.945. Three statistical charts were established to explain the prediction based on the characteristics of the XGboost model. Higher diagnostic performance was achieved with the XGboost model. These three statistical charts can help us understand why the machine learning model produces a specific prediction result. This may be the first attempt to apply “explainable artificial intelligence” to eye disease diagnosis.

Download Full-text

Development of a Machine Learning Model for the Prediction of the Real Time Mortality in Patients in the Intensive Care Unit

10.21203/rs.3.rs-1066192/v1 ◽

2021 ◽

Author(s):

Jaeyoung Yang ◽

Hong-Gook Lim ◽

Wonhyeong Park ◽

Dongseok Kim ◽

Jin Sun Yoon ◽

...

Keyword(s):

Machine Learning ◽

Intensive Care Unit ◽

Intensive Care ◽

Prediction Model ◽

Prediction Models ◽

External Validation ◽

Learning Model ◽

Mortality Prediction ◽

Machine Learning Model ◽

Mortality Prediction Model

Abstract BackgroundPrediction of mortality in intensive care units is very important. Thus, various mortality prediction models have been developed for this purpose. However, they do not accurately reflect the changing condition of the patient in real time. The aim of this study was to develop and evaluate a machine learning model that predicts short-term mortality in the intensive care unit using four easy-to-collect vital signs.MethodsTwo independent retrospective observational cohorts were included in this study. The primary training cohort included the data of 1968 patients admitted to the intensive care unit at the Veterans Health Service Medical Center, Seoul, South Korea, from January 2018 to March 2019. The external validation cohort comprised the records of 409 patients admitted to the medical intensive care unit at Seoul National University Hospital, Seoul, South Korea, from January 2019 to December 2019. Datasets of four vital signs (heart rate, systolic blood pressure, diastolic blood pressure, and peripheral capillary oxygen saturation [SpO2]) measured every hour for 10 h were used for the development of the machine learning model. The performances of mortality prediction models generated using five machine learning algorithms, Random Forest (RF), XGboost, perceptron, convolutional neural network, and Long Short-Term Memory, were calculated and compared using area under the receiver operating characteristic curve (AUROC) values and an external validation dataset.ResultsThe machine learning model generated using the RF algorithm showed the best performance. Its AUROC was 0.922, which is much better than the 0.8408 of the Acute Physiology and Chronic Health Evaluation II. Thus, to investigate the importance of variables that influence the performance of the machine learning model, machine learning models were generated for each observation time or vital sign using the RF algorithm. The machine learning model developed using SpO2 showed the best performance (AUROC, 0.89). ConclusionsThe mortality prediction model developed in this study using data from only four types of commonly recorded vital signs is simpler than any existing mortality prediction model. This simple yet powerful new mortality prediction model could be useful for early detection of probable mortality and appropriate medical intervention, especially in rapidly deteriorating patients.

Download Full-text

Machine Learning Model Based Expert System for Pig Disease Diagnosis

Communications in Computer and Information Science - Recent Trends in Image Processing and Pattern Recognition ◽

10.1007/978-981-16-0493-5_27 ◽

2021 ◽

pp. 302-312

Author(s):

Khumukcham Robindro ◽

Ksh. Nilakanta Singh ◽

Leishangthem Sashikumar Singh

Keyword(s):

Machine Learning ◽

Expert System ◽

Disease Diagnosis ◽

Learning Model ◽

Model Based ◽

Machine Learning Model

Download Full-text

MEWS++: Enhancing the Prediction of Clinical Deterioration in Admitted Patients through a Machine Learning Model

Journal of Clinical Medicine ◽

10.3390/jcm9020343 ◽

2020 ◽

Vol 9 (2) ◽

pp. 343 ◽

Cited By ~ 4

Author(s):

Arash Kia ◽

Prem Timsina ◽

Himanshu N. Joshi ◽

Eyal Klang ◽

Rohit R. Gupta ◽

...

Keyword(s):

Machine Learning ◽

At Risk ◽

Area Under The Curve ◽

Learning Model ◽

Clinical Deterioration ◽

Early Warning Score ◽

Support Vector ◽

Adult Age ◽

Machine Learning Model ◽

Patients At Risk

Early detection of patients at risk for clinical deterioration is crucial for timely intervention. Traditional detection systems rely on a limited set of variables and are unable to predict the time of decline. We describe a machine learning model called MEWS++ that enables the identification of patients at risk of escalation of care or death six hours prior to the event. A retrospective single-center cohort study was conducted from July 2011 to July 2017 of adult (age > 18) inpatients excluding psychiatric, parturient, and hospice patients. Three machine learning models were trained and tested: random forest (RF), linear support vector machine, and logistic regression. We compared the models’ performance to the traditional Modified Early Warning Score (MEWS) using sensitivity, specificity, and Area Under the Curve for Receiver Operating Characteristic (AUC-ROC) and Precision-Recall curves (AUC-PR). The primary outcome was escalation of care from a floor bed to an intensive care or step-down unit, or death, within 6 h. A total of 96,645 patients with 157,984 hospital encounters and 244,343 bed movements were included. Overall rate of escalation or death was 3.4%. The RF model had the best performance with sensitivity 81.6%, specificity 75.5%, AUC-ROC of 0.85, and AUC-PR of 0.37. Compared to traditional MEWS, sensitivity increased 37%, specificity increased 11%, and AUC-ROC increased 14%. This study found that using machine learning and readily available clinical data, clinical deterioration or death can be predicted 6 h prior to the event. The model we developed can warn of patient deterioration hours before the event, thus helping make timely clinical decisions.

Download Full-text

Identifying Factors Associated With Severe Intellectual Disabilities in Teenagers With Cerebral Palsy Using a Predictive Learning Model

Journal of Child Neurology ◽

10.1177/0883073818822358 ◽

2019 ◽

Vol 34 (4) ◽

pp. 221-229 ◽

Cited By ~ 6

Author(s):

Carlo M. Bertoncelli ◽

Paola Altamura ◽

Edgar Ramos Vieira ◽

Domenico Bertoncelli ◽

Susanne Thummler ◽

...

Keyword(s):

Machine Learning ◽

Cerebral Palsy ◽

Intellectual Disability ◽

Prediction Model ◽

Motor Skills ◽

Adaptive Functioning ◽

Learning Model ◽

Intelligence Scale ◽

Factors Associated ◽

Machine Learning Model

Background: Intellectual disability and impaired adaptive functioning are common in children with cerebral palsy, but there is a lack of studies assessing these issues in teenagers with cerebral palsy. Therefore, the aim of this study was to develop and test a predictive machine learning model to identify factors associated with intellectual disability in teenagers with cerebral palsy. Methods: This was a multicenter controlled cohort study of 91 teenagers with cerebral palsy (53 males, 38 females; mean age ± SD = 17 ± 1 y; range: 12-18 y). Data on etiology, diagnosis, spasticity, epilepsy, clinical history, communication abilities, behaviors, motor skills, eating, and drinking abilities were collected between 2005 and 2015. Intellectual disability was classified as “mild,” “moderate,” “severe,” or “profound” based on adaptive functioning, and according to the DSM-5 after 2013 and DSM-IV before 2013, the Wechsler Intelligence Scale for Children for patients up to ages 16 years, 11 months, and the Wechsler Adult Intelligence Scale for patients ages 17-18. Statistical analysis included Fisher’s exact test and multiple logistic regressions to identify factors associated with intellectual disability. A predictive machine learning model was developed to identify factors associated with having profound intellectual disability. The guidelines of the “Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis Statement” were followed. Results: Poor manual abilities ( P ≤ .001), gross motor function ( P ≤ .001), and type of epilepsy (intractable: P = .04; well controlled: P = .01) were significantly associated with profound intellectual disability. The average model accuracy, specificity, and sensitivity was 78%. Conclusion: Poor motor skills and epilepsy were associated with profound intellectual disability. The machine learning prediction model was able to adequately identify high likelihood of severe intellectual disability in teenagers with cerebral palsy.

Download Full-text

Covid-19 Analysis and Prediction using Data Science and Machine Learning

International Journal for Research in Applied Science and Engineering Technology ◽

10.22214/ijraset.2021.39272 ◽

2021 ◽

Vol 9 (12) ◽

pp. 303-307

Author(s):

Akshata Kulkarni

Keyword(s):

Machine Learning ◽

Data Science ◽

Information Dissemination ◽

Prediction Models ◽

Learning Model ◽

Future Trend ◽

Control Measures ◽

Machine Learning Model ◽

Using Data ◽

Novel Coronavirus

Abstract: Officials around the world are using several COVID-19 outbreak prediction models to make educated decisions and enact necessary control measures. In this study, we developed a Machine Learning model which predicts and forecasts the COVID-19 outbreak in India, with the goal of determining the best regression model for an in-depth examination of the novel coronavirus. Based on data available from January 31 to October 31, 2020, collected from Kaggle, this model predicts the number of confirmed cases in Maharashtra. We're using a Machine Learning model to foresee the future trend of these situations. The project has the potential to demonstrate the importance of information dissemination in improving response time and planning ahead of time to help reduce risk.

Download Full-text

Quantitative Toxicity Prediction via Ensembling of Heterogeneous Predictors

10.21203/rs.2.19338/v1 ◽

2019 ◽

Author(s):

Abdul Karim ◽

Vahid Riahi ◽

Avinash Mishra ◽

Abdollah Dehzangi ◽

M. A. Hakim Newton ◽

...

Keyword(s):

Machine Learning ◽

Deep Learning ◽

Prediction Models ◽

Individual Performance ◽

Learning Model ◽

Data Representation ◽

Toxicity Prediction ◽

Machine Learning Model ◽

Machine Learning Approach ◽

Benchmark Datasets

Abstract Representing molecules in the form of only one type of features and using those features to predict their activities is one of the most important approaches for machine-learning-based chemical-activity-prediction. For molecular activities like quantitative toxicity prediction, the performance depends on the type of features extracted and the machine learning approach used. For such cases, using one type of features and machine learning model restricts the prediction performance to specific representation and model used. In this paper, we study quantitative toxicity prediction and propose a machine learning model for the same. Our model uses an ensemble of heterogeneous predictors instead of typically using homogeneous predictors. The predictors that we use vary either on the type of features used or on the deep learning architecture employed. Each of these predictors presumably has its own strengths and weaknesses in terms of toxicity prediction. Our motivation is to make a combined model that utilizes different types of features and architectures to obtain better collective performance that could go beyond the performance of each individual predictor. We use six predictors in our model and test the model on four standard quantitative toxicity benchmark datasets. Experimental results show that our model outperforms the state-of-the-art toxicity prediction models in 8 out of 12 accuracy measures. Our experiments show that ensembling heterogeneous predictor improves the performance over single predictors and homogeneous ensembling of single predictors.The results show that each data representation or deep learning based predictor has its own strengths and weaknesses, thus employing a model ensembling multiple heterogeneous predictors could go beyond individual performance of each data representation or each predictor type.

Download Full-text

Preoperative prediction of complicated appendicitis using machine learning method

10.21203/rs.3.rs-27341/v1 ◽

2020 ◽

Author(s):

Chunbo Kang ◽

Xubin Li ◽

Xiaoqian Chi ◽

Yabin Yang ◽

Haifeng Shan ◽

...

Keyword(s):

Machine Learning ◽

Clinical Symptoms ◽

Laboratory Data ◽

Learning Model ◽

Complicated Appendicitis ◽

Support Vector ◽

Training Set ◽

Preoperative Prediction ◽

Machine Learning Model ◽

Testing Set

Abstract BACKGROUND Accurate preoperative prediction of complicated appendicitis (CA) could help selecting optimal treatment and reducing risks of postoperative complications. The study aimed to develop a machine learning model based on clinical symptoms and laboratory data for preoperatively predicting CA.METHODS 136 patients with clinicopathological diagnosis of acute appendicitis were retrospectively included in the study. The dataset was randomly divided (94: 42) into training and testing set. Predictive models using individual and combined selected clinical and laboratory data features were built separately. Three combined models were constructed using logistic regression (LR), support vector machine (SVM) and random forest (RF) algorithms. The CA prediction performance was evaluated with Receiver Operating Characteristic (ROC) analysis, using the area under the curve (AUC), sensitivity, specificity and accuracy factors.RESULTS The features of the abdominal pain time, nausea and vomiting, the highest temperature, high sensitivity-CRP (hs-CRP) and procalcitonin (PCT) had significant differences in the CA prediction (P<0.001). The ability to predict CA by individual feature was low (AUC<0.8). The prediction by combined features was significantly improved. The AUC of the three models (LR, SVM and RF) in the training set and the testing set were 0.805, 0.888, 0.908 and 0.794, 0.895, 0.761, respectively. The SVM-based model showed a better performance for CA prediction. RF had a higher AUC in the training set, but its poor efficiency in the testing set indicated a poor generalization ability.CONCLUSIONS The SVM machine learning model applying clinical and laboratory data can well predict CA preoperatively which could assist diagnosis in resource limited settings.

Download Full-text

An Intelligent Rockburst Prediction Model Based on Scorecard Methodology

Minerals ◽

10.3390/min11111294 ◽

2021 ◽

Vol 11 (11) ◽

pp. 1294

Author(s):

Honglei Wang ◽

Zhenlei Li ◽

Dazhao Song ◽

Xueqiu He ◽

Aleksei Sobolev ◽

...

Keyword(s):

Machine Learning ◽

Prediction Model ◽

False Alarm ◽

False Alarm Rate ◽

Learning Model ◽

Underground Engineering ◽

Risk Levels ◽

Model Based ◽

Machine Learning Model ◽

Sample Category

Rockburst is a serious hazard in underground engineering, and accurate prediction of rockburst risk is challenging. To construct an intelligent prediction model of rockburst risk with interpretability and high accuracy, three binary scorecards predicting different risk levels of rockburst were constructed using ChiMerge, evidence weight theory, and the logistic regression algorithm. An intelligent rockburst prediction model based on scorecard methodology (IRPSC) was obtained by integrating the three scorecards. The effects of hazard sample category weights on the missed alarm rate, false alarm rate, and accuracy of the IRPSC were analyzed. Results show that the accuracy, false alarm rate, and missed alarm rate of the IRPSC for rockburst prediction in riverside hydropower stations are 75%, 12.5%, and 12.5%, respectively. Setting higher hazard sample category weights can reduce the missed alarm rate of IRPSC, but it will lead to a higher false alarm rate. The IRPSC can adaptively adjust the threshold and weight value of the indicator and convert the abstract machine learning model into a tabular form, which overcomes the commonly black box problems of machine learning model, as well as is of great significance to the application of machine learning in rockburst risk prediction.

Download Full-text

Technical note: how to rationally compare the performances of different machine learning models?

10.7287/peerj.preprints.26714 ◽

2018 ◽

Cited By ~ 1

Author(s):

Terazima Maeda

Keyword(s):

Machine Learning ◽

Predictive Accuracy ◽

Learning Model ◽

Technical Note ◽

Learning Models ◽

Specific Data ◽

Machine Learning Model ◽

Specific Prediction ◽

Data Group ◽

Machine Learning Models

Nowadays, there is a large number of machine learning models that could be used for various areas. However, different research targets are usually sensitive to the type of models. For a specific prediction target, the predictive accuracy of a machine learning model is always dependent to the data feature, data size and the intrinsic relationship between inputs and outputs. Therefore, for a specific data group and a fixed prediction mission, how to rationally compare the predictive accuracy of different machine learning model is a big question. In this brief note, we show how should we compare the performances of different machine models by raising some typical examples.

Download Full-text

Moderate to severe OSA screening based on support vector machine of the Chinese population faciocervical measurements dataset: a cross-sectional study

BMJ Open ◽

10.1136/bmjopen-2020-048482 ◽

2021 ◽

Vol 11 (9) ◽

pp. e048482

Author(s):

Liu Zhang ◽

Ya Ru Yan ◽

Shi Qi Li ◽

Hong Peng Li ◽

Ying Ni Lin ◽

...

Keyword(s):

Machine Learning ◽

Chinese Population ◽

Learning Model ◽

Cross Sectional Study ◽

Neck Circumference ◽

Secondary Outcome ◽

Support Vector ◽

Sectional Study ◽

Cross Sectional ◽

Machine Learning Model

ObjectivesObstructive sleep apnoea (OSA) has received much attention as a risk factor for perioperative complications and 68.5% of OSA patients remain undiagnosed before surgery. Faciocervical characteristics may screen OSA for Asians due to smaller upper airways compared with Caucasians. Thus, our study aimed to explore a machine-learning model to screen moderate to severe OSA based on faciocervical and anthropometric measurements.DesignA cross-sectional study.SettingData were collected from the Shanghai Jiao Tong University School of Medicine affiliated Ruijin Hospital between February 2019 and August 2020.ParticipantsA total of 481 Chinese participants were included in the study.Primary and secondary outcome(1) Identification of moderate to severe OSA with apnoea–hypopnoea index 15 events/hour and (2) Verification of the machine-learning model.ResultsSex-Age-Body mass index (BMI)-maximum Interincisal distance-ratio of Height to thyrosternum distance-neck Circumference-waist Circumference (SABIHC2) model was set up. The SABIHC2 model could screen moderate to severe OSA with an area under the curve (AUC)=0.832, the sensitivity of 0.916 and specificity of 0.749, and performed better than the STOP-BANG (snoring, tiredness, observed apnea, high blood pressure, BMI, age, neck circumference, and male gender) questionnaire, which showed AUC=0.631, the sensitivity of 0.487 and specificity of 0.772. Especially for asymptomatic patients (Epworth Sleepiness Scale <10), the SABIHC2 model demonstrated better predictive ability compared with the STOP-BANG questionnaire, with AUC (0.824 vs 0.530), sensitivity (0.892 vs 0.348) and specificity (0.755 vs 0.809).ConclusionThe SABIHC2 machine-learning model provides a simple and accurate assessment of moderate to severe OSA in the Chinese population, especially for those without significant daytime sleepiness.

Download Full-text