Diabetes is a chronic disease that continues to be a primary and worldwide health concern since the health of the entire population has been affected by it. Over the years, many academics have attempted to develop a reliable diabetes prediction model using machine learning (ML) algorithms. However, these research investigations have had a minimal impact on clinical practice as the current studies focus mainly on improving the performance of complicated ML models while ignoring their explainability to clinical situations. Therefore, the physicians find it difficult to understand these models and rarely trust them for clinical use. In this study, a carefully constructed, efficient, and interpretable diabetes detection method using an explainable AI has been proposed. The Pima Indian diabetes dataset was used, containing a total of 768 instances where 268 are diabetic, and 500 cases are non-diabetic with several diabetic attributes. Here, six machine learning algorithms (artificial neural network (ANN), random forest (RF), support vector machine (SVM), logistic regression (LR), AdaBoost, XGBoost) have been used along with an ensemble classifier to diagnose the diabetes disease. For each machine learning model, global and local explanations have been produced using the Shapley additive explanations (SHAP), which are represented in different types of graphs to help physicians in understanding the model predictions. The balanced accuracy of the developed weighted ensemble model was 90% with a F1 score of 89% using a five-fold cross-validation (CV). The median values were used for the imputation of the missing values and the synthetic minority oversampling technique (SMOTETomek) was used to balance the classes of the dataset. The proposed approach can improve the clinical understanding of a diabetes diagnosis and help in taking necessary action at the very early stages of the disease.
31 March 2022 in Computational Biology and Chemistry
During the third wave of the coronavirus epidemic in Bangladesh, the death and infection rate due to this devastating virus has increased dramatically. The rapid spread of the virus is one of the reasons for this terrible condition. So, identifying the subsequent cases of coronavirus can be a great tool to reduce the mortality and infection rate. In this article, we used the autoregressive integrated moving average-ARIMA(8,1,7) model to estimate the expected daily number of COVID-19 cases in Bangladesh based on the data from April 20, 2021, to July 4, 2021. The ARIMA model showed the best results among the five executed models over Autoregressive Model (AR), Moving Average (MA), Autoregressive Moving Average (ARMA), and Rolling Forest Origin. The findings of this article were used to anticipate a rise in daily cases for the next month in Bangladesh, which can help governments plan policies to prevent the spread of the virus. The forecasting outcome indicated that this new trend(named delta variant) in Bangladesh would continue increasing and might reach 18327 daily new cases within four weeks if strict rules and regulations are not applied to control the spread of COVID-19.
Diabetic is a disease that occurred when the level of blood glucose is higher than usual, which is also known as hyperglycemia. When the human body is incapable of producing enough insulin(a hormone that produces glucose from food), then this situation leads to diabetes. The rapid increase of this disease makes the researchers work much harder in this area to build a model for diagnosing diabetes efficiently. As in healthcare, the availability of data is high, so it is easy to extract information from those data to diagnose disease and develop a new model for better results. This paper aims to introduce a model that can predict diabetes efficiently with the help of machine learning algorithms. Here logistic regression, SVM, and k nearest neighbor algorithms have been used for the classification of diabetics. After data preprocessing and training, those algorithms gave a good result. Logistic regression provided the best accuracy of 83% for test data. Also, SVM and knn both performed well and showed an accuracy of 82% and 79%, respectively. The proposed model has demonstrated improved results compared with previous work.
Autism Spectrum Disorder (ASD) is a genetic and neurological condition that leads to difficulties in communication and social interaction. The global concern associated with ASD diagnosis is increasing at a rapid rate due to its significant impact on quality of life. Early identification of ASD can significantly support timely intervention and treatment planning. While research in ASD diagnosis is evolving through the application of machine learning (ML) techniques, practical implementation in clinical settings has not progressed at the same pace. Although theoretical studies have demonstrated improved ML performance, they have not gained much interest among clinicians due to a lack of explainability. This study focuses on optimizing and comparing various machine learning models for ASD diagnosis, while incorporating explainable AI techniques to ensure model transparency and interpretability. The paper uses naive Bayes, Support Vector Machine (SVM), and Random Forest (RF) as classifiers after careful investigation. Three different publicly available datasets have been used based on the age group to create the best predicting model for each case. After handling missing values, balancing the dataset, and analyzing the classifier’s performance, it is found that tree-based algorithms, particularly RF, perform better for all the datasets. The RF model achieved up to 99% balanced accuracy on the adult dataset, with similarly strong performance on the children and adolescent datasets using five-fold cross-validation (CV). SHAP are also illustrated to improve model interpretability by highlighting the most influential features, thereby aiding physician understanding. The novelty of this work lies in the integration of explainable AI with robust preprocessing and age-specific modeling across multiple ASD datasets, addressing both diagnostic accuracy and clinical interpretability. It can be considered that the suggested method can efficiently diagnose ASD at a very early stage and enhance the understanding of ASD diagnosis clinically.
23 December 2024 in Neural Computing and Applications