In this study, we develop and compare quantum and classical machine learning-based chronic kidney disease prediction models. We used the "Chronic_Kidney_Disease Data Set" of the UCI Machine Learning Repository. We performed data preprocessing and applied feature engineering techniques to select the best features. We developed two quantum machine learning-based models and two classical machine learning-based models. We used a hybrid classical-quantum environment for building quantum machine learning models. Finally, we compared the performances of all four models. We found that the Quantum Support Vector Machine performs best among the quantum models. The model’s accuracy was 95% with a k-fold cross-validation score of 94.5% and an ROC-AUC score of 0.987. Among the classical models, the Support Vector Machine showed the highest performance with an accuracy of 92.5%, a k-fold cross-validation score of 93.9%, and a ROC-AUC score of 0.974. Overall, the Quantum Support Vector Machine outperformed all other developed models in terms of accuracy and validation scores. If the quantum models can be executed in a quantum computer instead of a hybrid environment, the models will exhibit higher accuracy and faster execution time. As the models are precisely predicting chronic kidney disease with high accuracy, we believe this study will act as an inspiring framework in the less-investigated field of quantum machine learning-based chronic kidney disease prediction.

Researchers are looking into Human Activity Recognition to find out different activities using affordable sensors. They are also using machine learning for real-time monitoring with sensor data. In healthcare, it's important to know how much energy is used during activities, especially for people who cannot move much, like older folks with Chronic Fatigue Syndrome. Figuring out their daily energy use is crucial to help them manage fatigue better. So, there's a need for a detailed study on finding different activities and understanding how much energy is used in every activity, especially in healthcare. In this survey paper, we explored articles about recognizing physical and social activities. We looked into the detailed methods used to calculate energy expenditure, carefully comparing the sensors used in various research projects. We investigated pre-processing techniques for sensor data. We uncovered how to extract features and studied the machine-learning methods used with raw sensor data. Bringing all this together not only broadened our understanding but also gave us a nuanced view of recognizing activities and calculating energy expenditure. This paper provides a detailed survey of recent sensor-based research focused on detecting and measuring physical activities. There is a scarcity of systems utilizing sensors for this purpose. As the number of individuals facing challenges in performing day-to-day activities continues to rise, there is an imperative need for enhanced methods to develop such systems.

Chronic Fatigue Syndrome (CFS) is a disorder with complex symptoms among patients. In most cases, CFS sufferers describe severe body weakness, poor sleep and inability to perform their usual work as their primary complaints. Symptoms worsen when the patient attempts to do similar work as tolerated. To prevent the worsening of symptoms, the patients need to be aware of what intensity of work they can manage. In this paper, we propose CFSCare, a hardware and software-based system that uses ML models to measure CFS patients' daily activity and energy expenditure objectively. Through our developed App, CFSCare submits to the user a summary of the comprehensive reports of the user's activity and sends a recommendation to the user on how they can prevent acquiring symptoms of CFS brought about by over-exertion. We use an Android smartphone and wrist sensor (MetamotionC) to monitor their leg and hand activity. We develop ML models based on SVM and DT algorithms to predict particular leg and hand activities. Among the applied ML models, SVM exhibited a brilliant performance with 98% accuracy in predicting leg activities and an average to-fold cross-validation score of 94%. For the hand activities prediction, DT recorded the best accuracy of 96%, and the average score of to cross-validations is also 96%. Since CFS patients can tire with exertion after a small amount of daily activity, CFSCare can playa vital role in preventing the patient from over-exertion through the monitoring features.

This paper introduces a novel concept of understanding cerebral dynamics by exploring the acoustic signals generated by blood flow in the brain (BFB) and mechanical resonant frequencies produced by the brain. The concept of this paper will be a groundbreaking approach to brain signal analysis through acoustic signals of BFB. Traditional methods for brain wave capture and analysis mostly depend on fMRI and EEG signals, which are now very popular and have some limitations in accessibility, cost, and real-time analysis capabilities. In this study, we seek the existing gaps in the current brain signal-capturing methods, and our theoretical underpinnings hypothesize that sound produced by blood flow in the brain (BFB) can be an innovative approach to capture the brain signal in a more user-friendly and accessible way. The feasibility of capturing and analyzing these BFB-sound is also discussed in this paper. We proposed a theoretical framework to capture the BFB sound through the human ear. The successful completion of this concept architecture will serve in different applications, from diagnosing neurological disorders to monitoring brain health, underscoring its potential to revolutionize non-invasive brain diagnostics. By synthesizing current knowledge and proposing innovative techniques, this paper aims to pave the way for new frontiers in understanding brain function through the brain sound generated by blood flow and captured from the human ears.

The measurement of Blood Urea Nitrogen (BUN) is an important test for examining the kidney condition. BUN is a waste product that is eliminated from the body by the kidneys. Increased BUN level often indicates renal diseases. The invasive blood test is used as the standard BUN testing method. In this paper, we have demonstrated an effective and innovative method for noninvasive BUN prediction from Ten-second fingertip video. We use the Ten-second fingertip video of human subjects to create photo plethysmography (PPG) features through signal processing and feature engineering techniques. We utilize Generative Adversarial Networks (GAN), a type of Generative AI, as a data augmentation method to create synthetic data based on our original data. Then, we use the PPG features and Gold standard BUN level of subj ects to develop four regression models before and after the implementation of GAN. We compare all models' performance in terms of MAE, MSE, and RMSE. Before applying GAN, ANN outperforms other models with an MAE of 1.17, MSE of 2.62, and RMSE of 1.62. We find that GAN implementation significantly improves every model's performance and provides lower MAE, MSE, and RMSE. After applying GAN, ANN performs best among all other models and shows the lowest MAE of 0.69, MSE of 1.03, and RMSE of 1.01. With these outstanding results, our noninvasive method of BUN prediction has the potential to offer enhanced patient comfort, early detection, continuous and remote monitoring of renal health.

In this paper, we introduce a machine learning (ML)-based approach for Chronic Kidney Disease (CKD) and diabetes prediction and perform feature effect analysis through SHAP (SHapley Additive exPlanations). We utilize two publicly available clinical datasets and build five ML classifier models for the analysis. Among all models, CatBoost provides the best performance for both CKD and diabetes prediction. Our CatBoost model has an accuracy of 0.95, mean 10-fold cross-validation of 0.96, ROC-AUC score of 0.99, precision of 0.96, recall of 0.96, and F1-score of 0.96 for CKD. The CatBoost model shows an accuracy of 0.99, mean 10-fold cross-validation of 0.97, ROC-AUC score of 1.0, precision of 1.0, recall of 0.98, and F1-score of 0.99 for diabetes. We perform comprehensive feature effect analysis by computing SHAP values. This gives insights into the contribution of every feature to the model's predictions. The achieved results highlight that CatBoost is a robust choice for accurate and reliable predictions of CKD and diabetes. The findings of this study contribute to the corresponding field of ML approaches for medical diagnosis and illustrate the significance of feature effect analysis in understanding model predictions. With excellent results, our research has the potential to enhance the clinical decision-making process and improve patient outcomes.

Natural Language Processing (NLP) is one of the fundamental domains of Artificial Intelligence (AI). In this paper, we present a systematic review of NLP based research for recognizing Bangla speech with regular and regional dialects. We describe 23 research papers based on Bangla speech recognition in regular accents and regional dialects. Due to the cultural diversity, Bangla has many dialects with distinctive regional pronunciations, complex vocabulary, and syntax. These characteristics of Bangla create several challenges for implementing NLP successfully. In this paper, we focus on NLP's vital role in speech recognition, which is essential to virtual assistants, transcription services, and language-learning applications. We discuss several methods such as advanced language models, comprehensive datasets, and continuous adaptation to overcome the challenges of recognizing Bangla with NLP. Several algorithms, such as Deep Neural Networks, Gaussian Mixture Models (GMM), Linear Predictive Coding (LPC), Mel frequency cepstral coefficients (MFCC), etc. have been developed to detect the spoken words in Bangla. These existing research works have faced various challenges like scarcity of data, dialectal variability, computational resource requirements, etc. For mitigation of the challenges and further advancement in this research area, we discuss some research scopes including dialect identification through large datasets, low-resource dialect modeling, development of deep learning model for end-to-end ASR systems, continuous learning, etc. This research will help us understand the present status and challenges associated with NLP-based Bangla speech recognition.

The aim of this study is to enhance the accuracy and reduce the time complexity of predicting cardiac illnesses by utilizing patient vital signs exclusively. The importance of timely treatment in critical situations where quick decisions are necessary is emphasized here. Precise predictions have the potential to prevent health deterioration and even save lives. By analyzing vast datasets (images, texts etc.) and detecting subtle patterns, quantum machine learning algorithms (QML) can offer more accurate results in a shorter period compared to classical algorithms. The feature set used here is a novel one to be used for the quick detection and early prediction of cardiovascular diseases, and also a convenient one for this task. To predict heart diseases or abnormalities, both classical and quantum machine learning techniques have been employed in this paper. We have developed four models based on Support Vector Machine (SVM), Neural Network (NN), Quantum Support Vector Machine (QSVM), and Quantum Neural Network (QNN) and compared their performance on a balanced sample of a vast dataset for heart diseases prediction. After comparing the models' performance, we found that Quantum Support Vector Machine (QSVM) performed best with an accuracy of 75% and to-fold Cross-validation score of 80%.

The paper presents chaos-theory-based human attention level detection from the electroencephalogram (EEG) signals. In the medical field, “human attention level” can be referred to as the “attentional state,”, which helps to understand an individual's attention capacity in various crucial moments. In this study, we have deployed secondary analysis on existing methods by implementing chaos theory on the PhysioNet dataset. We investigate different vital parameters and values to predict human attention level from the EEG dataset. We calculated time delay, embedding dimension, and correlation dimension from the participants' EEG data to determine the parameters' values: specifically, for detecting of human attention level from EEG signals. By calculating the 95% confidence interval (CI), the time delay has an average of 2.50 seconds, and the embedding dimension and correlation dimension have an average value of 4.41 and 2.23, respectively. We also observed a similar embedded signal in the reconstructed phase space (RPS) of participant's EEG signals. The statistical and chaos-based plot can potentially investigate human attention parameters and develop a robust EEG signal prediction system. Overall, this proposed framework serves as a resource on the latest nonlinearity detection techniques to detect human attention levels utilizing EEG signal analysis. Clinical Relevance - The effectiveness of the chaos-based non-linear analysis method for detecting human attention levels in EEG signals depends on its potential impact on diagnosis and treatment, integration into clinical practice, benefits, and risks.

The paper presents the design methodology and implementation of a low-cost spirometer prototype that can detect pulmonary diseases. By sensing the pressure difference of the expiratory air passing through the 3D printed venturi tube, the system calculates forced vital capacity (FVC). Through free campaigns, the performance and efficiency of this venturi tube spirometer were tested in different locations and the accuracy rate was highly promising. With this great efficiency and accuracy, this model diagnoses chronic obstructive pulmonary disease (COPD). Apart from measuring FVC, the model also measures forced expiratory volume in one second (FEV1) and from their ratio, it identifies COPD. In underprivileged areas, each and every hospital cannot afford spirometers available in the market because of their high price. This low-cost spirometer model ensures proper detection of pulmonary diseases in those underprivileged areas and it is definitely going to add a new dimension to the diagnosis of COPD.