All Papers

Md. Kamrul Hasan, Md. Ashraful Alam, Dola Das, Eklas Hossain et al.

Diabetes, also known as chronic illness, is a group of metabolic diseases due to a high level of sugar in the blood over a long period. The risk factor and severity of diabetes can be reduced significantly if the precise early prediction is possible. The robust and accurate prediction of diabetes is highly challenging due to the limited number of labeled data and also the presence of outliers (or missing values) in the diabetes datasets. In this literature, we are proposing a robust framework for diabetes prediction where the outlier rejection, filling the missing values, data standardization, feature selection, K-fold cross-validation, and different Machine Learning (ML) classifiers (k-nearest Neighbour, Decision Trees, Random Forest, AdaBoost, Naive Bayes, and XGBoost) and Multilayer Perceptron (MLP) were employed. The weighted ensembling of different ML models is also proposed, in this literature, to improve the prediction of diabetes where the weights are estimated from the corresponding Area Under ROC Curve (AUC) of the ML model. AUC is chosen as the performance metric, which is then maximized during hyperparameter tuning using the grid search technique. All the experiments, in this literature, were conducted under the same experimental conditions using the Pima Indian Diabetes Dataset. From all the extensive experiments, our proposed ensembling classifier is the best performing classifier with the sensitivity, specificity, false omission rate, diagnostic odds ratio, and AUC as 0.789, 0.934, 0.092, 66.234, and 0.950 respectively which outperforms the state-of-the-art results by 2.00 % in AUC. Our proposed framework for the diabetes prediction outperforms the other methods discussed in the article. It can also provide better results on the same dataset which can lead to better performance in diabetes prediction. Our source code for diabetes prediction is made publicly available.

Pronab Ghosh, Sami Azam, Mirjam Jonkman, Asif Karim et al.

Cardiovascular diseases (CVD) are among the most common serious illnesses affecting human health. CVDs may be prevented or mitigated by early diagnosis, and this may reduce mortality rates. Identifying risk factors using machine learning models is a promising approach. We would like to propose a model that incorporates different methods to achieve effective prediction of heart disease. For our proposed model to be successful, we have used efficient Data Collection, Data Pre-processing and Data Transformation methods to create accurate information for the training model. We have used a combined dataset (Cleveland, Long Beach VA, Switzerland, Hungarian and Stat log). Suitable features are selected by using the Relief, and Least Absolute Shrinkage and Selection Operator (LASSO) techniques. New hybrid classifiers like Decision Tree Bagging Method (DTBM), Random Forest Bagging Method (RFBM), K-Nearest Neighbors Bagging Method (KNNBM), AdaBoost Boosting Method (ABBM), and Gradient Boosting Boosting Method (GBBM) are developed by integrating the traditional classifiers with bagging and boosting methods, which are used in the training process. We have also instrumented some machine learning algorithms to calculate the Accuracy (ACC), Sensitivity (SEN), Error Rate, Precision (PRE) and F1 Score (F1) of our model, along with the Negative Predictive Value (NPR), False Positive Rate (FPR), and False Negative Rate (FNR). The results are shown separately to provide comparisons. Based on the result analysis, we can conclude that our proposed model produced the highest accuracy while using RFBM and Relief feature selection methods (99.05%).

Mohammad Shafenoor Amin, Yin Kia Chiam, Kasturi Dewi Varathan

Md. Mamun Ali, Bikash Kumar Paul, Kawsar Ahmed, Francis M. Bui et al.

Machine learning and data mining-based approaches to prediction and detection of heart disease would be of great clinical utility, but are highly challenging to develop. In most countries there is a lack of cardiovascular expertise and a significant rate of incorrectly diagnosed cases which could be addressed by developing accurate and efficient early-stage heart disease prediction by analytical support of clinical decision-making with digital patient records. This study aimed to identify machine learning classifiers with the highest accuracy for such diagnostic purposes. Several supervised machine-learning algorithms were applied and compared for performance and accuracy in heart disease prediction. Feature importance scores for each feature were estimated for all applied algorithms except MLP and KNN. All the features were ranked based on the importance score to find those giving high heart disease predictions. This study found that using a heart disease dataset collected from Kaggle three-classification based on k-nearest neighbor (KNN), decision tree (DT) and random forests (RF) algorithms the RF method achieved 100% accuracy along with 100% sensitivity and specificity. Thus, we found that a relatively simple supervised machine learning algorithm can be used to make heart disease predictions with very high accuracy and excellent potential utility.

RichardB. Habershon

Md. Maniruzzaman, Md. Jahanur Rahman, Benojir Ahammed, Md. Menhazul Abedin

Background and objectives Diabetes is a chronic disease characterized by high blood sugar. It may cause many complicated disease like stroke, kidney failure, heart attack, etc. About 422 million people were affected by diabetes disease in worldwide in 2014. The figure will be reached 642 million in 2040. The main objective of this study is to develop a machine learning (ML)-based system for predicting diabetic patients. Materials and methods Logistic regression (LR) is used to identify the risk factors for diabetes disease based on p value and odds ratio (OR). We have adopted four classifiers like naïve Bayes (NB), decision tree (DT), Adaboost (AB), and random forest (RF) to predict the diabetic patients. Three types of partition protocols (K2, K5, and K10) have also adopted and repeated these protocols into 20 trails. Performances of these classifiers are evaluated using accuracy (ACC) and area under the curve (AUC). Results We have used diabetes dataset, conducted in 2009-2012, derived from the National Health and Nutrition Examination Survey. The dataset consists of 6561 respondents with 657 diabetic and 5904 controls. LR model demonstrates that 7 factors out of 14 as age, education, BMI, systolic BP, diastolic BP, direct cholesterol, and total cholesterol are the risk factors for diabetes. The overall ACC of ML-based system is 90.62% . The combination of LR-based feature selection and RF-based classifier gives 94.25% ACC and 0.95 AUC for K10 protocol. Conclusion The combination of LR and RF-based classifier performs better. This combination will be very helpful for predicting diabetic patients.

Isfafuzzaman Tasin, Tansin Ullah Nabil, Sanjida Islam, Riasat Khan

Globally, diabetes affects 537 million people, making it the deadliest and the most common non-communicable disease. Many factors can cause a person to get affected by diabetes, like excessive body weight, abnormal cholesterol level, family history, physical inactivity, bad food habit etc. Increased urination is one of the most common symptoms of this disease. People with diabetes for a long time can get several complications like heart disorder, kidney disease, nerve damage, diabetic retinopathy etc. But its risk can be reduced if it is predicted early. In this paper, an automatic diabetes prediction system has been developed using a private dataset of female patients in Bangladesh and various machine learning techniques. The authors used the Pima Indian diabetes dataset and collected additional samples from 203 individuals from a local textile factory in Bangladesh. Feature selection algorithm mutual information has been applied in this work. A semi-supervised model with extreme gradient boosting has been utilized to predict the insulin features of the private dataset. SMOTE and ADASYN approaches have been employed to manage the class imbalance problem. The authors used machine learning classification methods, that is, decision tree, SVM, Random Forest, Logistic Regression, KNN, and various ensemble techniques, to determine which algorithm produces the best prediction results. After training on and testing all the classification models, the proposed system provided the best result in the XGBoost classifier with the ADASYN approach with 81% accuracy, 0.81 F1 coefficient and AUC of 0.84. Furthermore, the domain adaptation method has been implemented to demonstrate the versatility of the proposed system. The explainable AI approach with LIME and SHAP frameworks is implemented to understand how the model predicts the final results. Finally, a website framework and an Android smartphone application have been developed to input various features and predict diabetes instantaneously. The private dataset of female Bangladeshi patients and programming codes are available at the following link: https://github.com/tansin-nabil/Diabetes-Prediction-Using-Machine-Learning.

S Campbell, Marion Campbell, Jeremy Grimshaw, Alison Walker

BACKGROUND: The aim of the study was to review systematically the literature measuring the accuracy of routine UK hospital statistics that classify patients on discharge. METHODS: A systematic review was carried out of studies comparing routine discharge statistics about an episode of hospital care with the original medical record. Dual quality assessment and extraction was completed for included studies. Qualitative and descriptive analyses were undertaken. Additional comparisons of factors that could potentially introduce systematic variation in coding accuracy were also undertaken. RESULTS: Thirty studies were identified, of which 21 were included in the review. Twelve of these were conducted in England and Wales, and nine in Scotland. The majority assessed the accuracy of a single diagnosis, or selection of diagnoses in a limited range of hospital settings. The median coding accuracy rates were 91 per cent for diagnostic codes and 69.5 per cent for operation or procedure codes in studies in England or Wales; 82 per cent for diagnostic codes and 98 per cent for operation or procedure codes in Scottish studies. There were no significant differences in coding accuracy over time or in the type or rarity of the codes being assessed. Accuracy rates were higher for ICD7 codes (median 96.5 per cent) than for ICD8 (median 87 per cent) or ICD9 (median 77 per cent). CONCLUSIONS: Coding accuracy on average is high in the United Kingdom, especially for operations and procedures. However, policy-makers, planners and researchers need to recognize and account for the degree of inaccuracy in routine hospital information statistics. Further research is needed into methods of improving and maintaining coding accuracy.

Md. Ariful Islam, Md. Ziaul Hasan Majumder, Md. Alomgeer Hussein

Chronic kidney disease (CKD) is a dangerous ailment that can last a person's entire life and is caused by either kidney malignancy or decreased kidney functioning. It is feasible to halt or slow the progression of this chronic disease to an end-stage wherein dialysis or surgical intervention is the only method to preserve a patient's life. Earlier detection and appropriate therapy can increase the likelihood of this happening. Throughout this research, the potential of several different machine learning approaches for providing an early diagnosis of CKD has been investigated. There has been a significant amount of research conducted on this topic. Nevertheless, we are bolstering our approach by making use of predictive modeling. Therefore, in our approach, we investigate the link that exists between data factors as well as the characteristics of the target class. We are capable of constructing a collection of prediction models with the help of machine learning and predictive analytics, thanks to the better measures of attributes that can be introduced using predictive modeling. This study starts with 25 variables in addition to the class property, but by the end, it has narrowed the list down to 30% of those parameters as the best subset to identify CKD. Twelve different machine learning-based classifiers have been tested in a supervised learning environment. Within the confines of a supervised learning environment, a total of 12 different machine learning-based classifiers have indeed been examined, with the greatest performance indicators being an accuracy of 0.983, a precision of 0.98, a recall of 0.98, and an F1-score of 0.98 for the XgBoost classifier. The way the research was done leads to the conclusion that recent improvements in machine learning, along with the help of predictive modeling, make for an interesting way to find new solutions that can then be used to test the accuracy of prediction in the field of kidney disease and beyond.

Md. Maniruzzaman, Md. Jahanur Rahman, Md. Al-MehediHasan, Harman S. Suri et al.

Diabetes mellitus is a group of metabolic diseases in which blood sugar levels are too high. About 8.8% of the world was diabetic in 2017. It is projected that this will reach nearly 10% by 2045. The major challenge is that when machine learning-based classifiers are applied to such data sets for risk stratification, leads to lower performance. Thus, our objective is to develop an optimized and robust machine learning (ML) system under the assumption that missing values or outliers if replaced by a median configuration will yield higher risk stratification accuracy. This ML-based risk stratification is designed, optimized and evaluated, where: (i) the features are extracted and optimized from the six feature selection techniques (random forest, logistic regression, mutual information, principal component analysis, analysis of variance, and Fisher discriminant ratio) and combined with ten different types of classifiers (linear discriminant analysis, quadratic discriminant analysis, naïve Bayes, Gaussian process classification, support vector machine, artificial neural network, Adaboost, logistic regression, decision tree, and random forest) under the hypothesis that both missing values and outliers when replaced by computed medians will improve the risk stratification accuracy. Pima Indian diabetic dataset (768 patients: 268 diabetic and 500 controls) was used. Our results demonstrate that on replacing the missing values and outliers by group median and median values, respectively and further using the combination of random forest feature selection and random forest classification technique yields an accuracy, sensitivity, specificity, positive predictive value, negative predictive value and area under the curve as: 92.26%, 95.96%, 79.72%, 91.14%, 91.20%, and 0.93, respectively. This is an improvement of 10% over previously developed techniques published in literature. The system was validated for its stability and reliability. RF-based model showed the best performance when outliers are replaced by median values.

Md. Maniruzzaman, Nishith Kumar, Md. Menhazul Abedin, Md. Shaykhul Islam et al.

Background and objective Diabetes is a silent killer. The main cause of this disease is the presence of excessive amounts of metabolites such as glucose. There were about 387 million diabetic people all over the world in 2014. The financial burden of this disease has been calculated to be about $13,700 per year. According to the World Health Organization (WHO), these figures will more than double by the year 2030. This cost will be reduced dramatically if someone can predict diabetes statistically on the basis of some covariates. Although several classification techniques are available, it is very difficult to classify diabetes. The main objectives of this paper are as follows: (i) Gaussian process classification (GPC), (ii) comparative classifier for diabetes data classification, (iii) data analysis using the cross-validation approach, (iv) interpretation of the data analysis and (v) benchmarking our method against others. Methods To classify diabetes, several classification techniques are used such as linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and Naive Bayes (NB). However, most of the medical data show non-normality, non-linearity and inherent correlation structure. So in this paper we adapted Gaussian process (GP)-based classification technique using three kernels namely: linear, polynomial and radial basis kernel. We also investigate the performance of a GP-based classification technique in comparison to existing techniques such as LDA, QDA and NB. Performances are evaluated by using the accuracy (ACC), sensitivity (SE), specificity (SP), positive predictive value (PPV), negative predictive value (NPV) and receiver-operating characteristic (ROC) curves. Results Pima Indian diabetes dataset is taken as part of the study. This consists of 768 patients, of which 268 patients are diabetic and 500 patients are controls. Our machine learning system shows the performance of GP-based model as: ACC 81.97%, SE 91.79%, SP 63.33%, PPV 84.91% and NPV 62.50% which are larger compared to other methods.

Safial Islam Ayon, Md. Milon Islam, Md Rahat Hossain

Diseases is an unusual circumstance that affects single or more parts of a human’s body. Because of lifestyle and patrimonial, different kinds of disease are increasing day by day. Among all those diseases, heart disease turns out to be the most common disease and the impact of this ailment is dangerous than all other diseases. In this paper, we compared a number of computational intelligence techniques for the prediction of coronary artery heart disease. Seven computational intelligence techniques named as Logistic Regression (LR), Support Vector Machine (SVM), Deep Neural Network (DNN), Decision Tree (DT), Naïve Bayes (NB), Random Forest (RF), and K-Nearest Neighbor (K-NN) were applied and a comparative study was drawn. The performance of each technique was evaluated using Statlog and Cleveland heart disease dataset which are retrieved from the UCI machine learning repository database with several evaluation techniques. From the study, it can be carried out that the highest accuracy of 98.15% obtained by deep neural network with sensitivity and precision 98.67% and 98.01% respectively. The outcomes of the study were compared with the outcomes of the state of the art focusing on heart disease prediction that outperforms the previous study.

Md. Manowarul Islam, Rahatara Ferdousi, Sadikur Rahman, Humayra Yasmin Bushra

Md. Zahangir Alam, Md. Saifur Rahman, M. Sohel Rahman

Medical data classification is considered to be a challenging task in the field of medical informatics. Although many works have been reported in the literature, there is still scope for improvement. In this paper, a feature ranking based approach is developed and implemented for medical data classification. The features of a dataset are ranked using some suitable ranker algorithms, and subsequently the Random Forest classifier is applied only on highly ranked features to construct the predictor. We have conducted extensive experiments on 10 benchmark datasets and the results are promising. We present highly accurate predictors for 10 different diseases, as well as suggest a methodology that is sufficiently general and is expected to perform well for other diseases with similar datasets.

Shadman Nashif, Md. Rakib Raihan, Md. Rasedul Islam, Mohammad Hasan Imam

Cardiovascular diseases are the most common cause of death worldwide over the last few decades in the developed as well as underdeveloped and developing countries. Early detection of cardiac diseases and continuous supervision of clinicians can reduce the mortality rate. However, accurate detection of heart diseases in all cases and consultation of a patient for 24 hours by a doctor is not available since it requires more sapience, time and expertise. In this study, a tentative design of a cloud-based heart disease prediction system had been proposed to detect impending heart disease using Machine learning techniques. For the accurate detection of the heart disease, an efficient machine learning technique should be used which had been derived from a distinctive analysis among several machine learning algorithms in a Java Based Open Access Data Mining Platform, WEKA. The proposed algorithm was validated using two widely used open-access database, where 10-fold cross-validation is applied in order to analyze the performance of heart disease detection. An accuracy level of 97.53% accuracy was found from the SVM algorithm along with sensitivity and specificity of 97.50% and 94.94%respectively. Moreover, to monitor the heart disease patient round-the-clock by his/her caretaker/doctor, a real-time patient monitoring system was developed and presented using Arduino, capable of sensing some real-time parameters such as body temperature, blood pressure, humidity, heartbeat. The developed system can transmit the recorded data to a central server which are updated every 10 seconds. As a result, the doctors can visualize the patient’s real-time sensor data by using the application and start live video streaming if instant medication is required. Another important feature of the proposed system was that as soon as any real-time parameter of the patient exceeds the threshold, the prescribed doctor is notified at once through GSM technology.

Umarani Nagavelli, Debabrata Samanta, Partha Chakraborty

At present, a multifaceted clinical disease known as heart failure disease can affect a greater number of people in the world. In the early stages, to evaluate and diagnose the disease of heart failure, cardiac centers and hospitals are heavily based on ECG. The ECG can be considered as a regular tool. Heart disease early detection is a critical concern in healthcare services (HCS). This paper presents the different machine learning technologies based on heart disease detection brief analysis. Firstly, Naïve Bayes with a weighted approach is used for predicting heart disease. The second one, according to the features of frequency domain, time domain, and information theory, is automatic and analyze ischemic heart disease localization/detection. Two classifiers such as support vector machine (SVM) with XGBoost with the best performance are selected for the classification in this method. The third one is the heart failure automatic identification method by using an improved SVM based on the duality optimization scheme also analyzed. Finally, for a clinical decision support system (CDSS), an effective heart disease prediction model (HDPM) is used, which includes density-based spatial clustering of applications with noise (DBSCAN) for outlier detection and elimination, a hybrid synthetic minority over-sampling technique-edited nearest neighbor (SMOTE-ENN) for balancing the training data distribution, and XGBoost for heart disease prediction. Machine learning can be applied in the medical industry for disease diagnosis, detection, and prediction. The major purpose of this paper is to give clinicians a tool to help them diagnose heart problems early on. As a result, it will be easier to treat patients effectively and avoid serious repercussions. This study uses XGBoost to test alternative decision tree classification algorithms in the hopes of improving the accuracy of heart disease diagnosis. In terms of precision, accuracy, f1-measure, and recall as performance parameters above mentioned, four types of machine learning (ML) models are compared.

Tuhin Biswas, Sarah P. Garnett, Sonia Pervin, Lal Rawal

BACKGROUND: Over the two last decades Bangladesh, a low-income country, has experienced a rapid demographic and epidemiological transition. The population has increased substantially with rapid urbanization and changing pattern of disease, which at least in part, can be explained by nutritional changes. However, the nutritional status of the adult population has not been previously described. Hence, the objective of this study was to estimate the prevalence and explore socio-demographic determinants of underweight, overweight and obesity among the Bangladeshi adult population. METHODS: This study is a secondary data analysis of the national 2011 Bangladesh Demographic and Health Survey. We determined the nutritional status of adults aged ≥35 years of age, who had a measured weight and height, using the Asian body mass index (BMI) cut-offs for underweight (BMI <18.5 kg/m2), overweight (BMI 23 to <27.5 kg/m2) and obesity (BMI ≥27.5 kg/m2). Logistic regression modeling was used to determine the association between socio-demographic factors and nutritional status. RESULT: Of total sample (n = 5495), 30.4% were underweight, 18.9% were overweight and 4.6% were obese. Underweight was associated with age, education and wealth. The adjusted odd ratios for underweight were higher for older people (≥70 years) compared to younger, the least educated compared to the most educated and the poorest compared to the wealthiest were 2.51 (95%CI: 1.95-3.23, p<0.001), 3.59 (95%CI: 2.30-5.61, p<0.001) and 3.70 (95%CI: 2.76-4.96, p<0.001), respectively. Younger age (35-44 years), being female, higher education, wealthier and living in urban areas were associated with overweight/obesity with adjusted odds ratios of 1.73 (95%CI: 1.24-2.41, p<0.001), 2.48 (95%CI: 1.87-3.28, p<0.001), 3.98 (95%CI: 2.96-5.33, p<0.001), 7.14 (95%CI: 5.20-9.81, p<0.001) 1.27 (95%CI: 1.05-1.55, p-0.02), respectively. CONCLUSION: Underweight and overweight/obesity are prevalent in Bangladeshi adults. Both conditions are associated with increased morbidity and mortality and increase the risk of developing non-communicable diseases. Effective public health intervention approaches are necessary to address both these conditions.

Zunaid Ahsan Karar, Nurul Alam, Peter Kim Streatfield

BACKGROUND: For understanding epidemiological transition, Health and Demographic Surveillance System plays an important role in developing and resource-constraint setup where accurate information on vital events (e.g. births, deaths) and cause of death is not available. METHODS: This study aimed to assess existing level and trend of causes of 18,917 deaths in Matlab, a rural area of Bangladesh, during 1986-2006 and to project future scenarios for selected major causes of death. RESULTS: The results demonstrated that Matlab experienced a massive change in the mortality profile from acute, infectious, and parasitic diseases to non-communicable, degenerative, and chronic diseases during the last 20 years. It also showed that over the period 1986-2006, age-standardized mortality rate (for both sexes) due to diarrhea and dysentery reduced by 86%, respiratory infections by 79%, except for tuberculosis which increased by 173%. On the other hand, during the same period, mortality due to cardiovascular and cerebrovascular diseases increased by a massive 3,527% and malignant neoplasms by 495%, whereas mortality due to chronic obstructive pulmonary disease and injury remained in the similar level (12-13% increase). CONCLUSION: The trend of selected causes of death demonstrates that in next two decades, deaths due to communicable diseases will decline substantially and the mortality due to non-communicable diseases (NCDs) will increase at massive proportions. Despite Matlab's significant advances in socio-demographic indicators, emergence of NCDs and mortality associated with it would be the major cause for concern in the coming years.

Safial Islam Ayon, Md. Milon Islam

Nowadays, Diabetes is one of the most common and severe diseases in Bangladesh as well as all over the world. It is not only harmful to the blood but also causes different kinds of diseases like blindness, renal disease, kidney problem, heart diseases etc. that causes a lot of death per year. So, it badly needs to develop a system that can effectively diagnose the diabetes patients using medical details. We propose a strategy for the diagnosis of diabetes using deep neural network by training its attributes in five-fold and ten-fold crossvalidation fashion. The Pima Indian Diabetes (PID) data set is retrieved from the UCI machine learning repository database. The results on PID dataset demonstrate that deep learning approach design an auspicious system for the prediction of diabetes with prediction accuracy of 98.35%, F1 score of 98, and MCC of 97 for five-fold cross-validation. Additionally, accuracy of 97.11%, sensitivity of 96.25%, and specificity of 98.80% are obtained for ten-fold cross-validation. The experimental results exhibit that the proposed system provides promising results in case of five-fold cross-validation.