Supporting Hypothesis Generation by Machine Learning in Smart Health

This study employs the framework of Web services in conjunction with Bayesian theorem and decision trees to construct a Web-services-based decision support system for medical diagnosis and treatment. The purpose is to help users (physicians) with issues pertinent to medical diagnosis and treatment decisions. Users through the system key in available prior probability and through computation based on Bayesian theorem obtain the diagnosis. The process helps users enhance the quality and efficiency of medical decisions, and the diagnosis can be transmitted to a decision-tree-based treatment decision support service component via XML to generate recommendation and analysis for treatment decisions. On the other hand, features of web services enable this medical decision support system to offer more service platforms than conventional one. Users will have access to the system whether they use Windows, Macintosh, Linux or any other platforms that connect with the Internet via HTTP. More important is the fact that after the system is completed all Internet service providers will be able to access the system as a software unit freely and quickly. This way, the goal of this study to provide medical decision support tools and speedily integrate heterogeneous medical decision support systems can be effectively attained.

Machine learning (ML) applications in medical artificial intelligence (AI) systems have shifted from traditional and statistical methods to increasing application of deep learning models. This survey navigates the current landscape of multimodal ML, focusing on its profound impact on medical image analysis and clinical decision support systems. Emphasizing challenges and innovations in addressing multimodal representation, fusion, translation, alignment, and co-learning, the paper explores the transformative potential of multimodal models for clinical predictions. It also highlights the need for principled assessments and practical implementation of such models, bringing attention to the dynamics between decision support systems and healthcare providers and personnel. Despite advancements, challenges such as data biases and the scarcity of “big data” in many biomedical domains persist. We conclude with a discussion on principled innovation and collaborative efforts to further the mission of seamless integration of multimodal ML models into biomedical practice.

Neural network modeling for surgical decisions on traumatic brain injury patients

The implementation of the medical decision support system (MDSS) in clinical practice requires careful monitoring to ensure patient safety and track the performance of artificial intelligence technologies.Goal: To assess the effectiveness of the TOP-3 MDSS in outpatient clinics of the Moscow Heath Care Department.Materials and methods: The Moscow Heath Care Department monitored the TOP-3 operation between October 1, 2020 and March 03, 2024 (n = 63,809,360 people). The Hit-3 metric was used to determine if the MDSS needs retraining. An additional study involving medical experts included a retrospective analysis of data from 3,000 patients and calculation of the diagnostic agreement rate among the MDSS, doctors and the experts.Results: The monitoring demonstrated mean Hit-3 of 63.5, 64.5 and 67.7 for MDSS ver. 1, 2, and 3, respectively. In cases where the clinicians disagreed with the MDSS (n = 2000), based on the patient complaints, experts agreed with the MDSS in 80.2% of cases, with clinicians in 11.5% of cases, and established a different diagnosis in 8.3% of cases. In cases where the clinicians’ and the MDSS’s conclusions matched (n = 1000), experts approved the diagnosis in 50.4% of cases, selected an alternative diagnosis suggested by the MDSS in 37.9% of cases, and established a different diagnosis for 11.7% of patients. Conclusion: The described monitoring methodology, supplemented by expert review, allowed for a comprehensive assessment of the MDSS to be implemented in the health care system. Based on the results of the TOP-3 effectiveness assessment, it was decided that the analyzed list of data from electronic health records should be expanded, which will be implemented in the upcoming version TOP-3+.

The enormous volume of medical data has resulted in the development of sophisticated systems that facilitate information search and enable clinicians in decision-making process. Driven by artificial intelligence, knowledge graphs (KGs) provide a solid structure for organising and evaluating vast volumes of diverse medical data, therefore enabling wiser question development and improved decision-making. This article presents a whole strategy for integrating knowledge graphs with artificial intelligence-based approaches to improve medical information search and decision support systems performance. Graph-based reasoning, natural language processing (NLP), and machine learning all help the proposed approach to enhance semantic comprehension. It achieves this by tying together unorganised and organised medical data sources to provide pertinent analysis. Using predictive analytics, personalised healthcare recommendations, and real-time clinical decision support, the AI-powered knowledge graph architecture helps you It achieves this by continuously shifting the relationships among illnesses, symptoms, therapies, pasts of patients. This approach also ensures that many healthcare systems may cooperate better, which facilitates information search and reduces the diagnostic error count. Including reinforcement learning techniques enhances question results depending on user interaction, therefore enhancing the search process. The results of experiments show that KGs with AI work better than traditional database-driven methods when it comes to getting medical information quickly, correctly, and usefully. The suggested method helps healthcare workers a lot by making it easier for them to get accurate, evidence-based information more quickly. This will eventually lead to better patient results. This study shows that knowledge graphs driven by AI have the ability to completely change how medical information is managed and how decisions are made. This could lead to smarter and more flexible healthcare systems.

Background and Aims Some of the most common complications of CKD are anemia and bone and mineral disorders (BMD). Adequacy of treatment of patients with these complications directly affects their life quality and duration. The goal of our work is to create medical decision support system (MDSS) that will be able to generate recommendations on diagnostics and treatment of CKD patients with anemia and BMD complications, based on clinical guidelines and control the fulfillment of corresponding prescriptions. Method We have developed an automated medical decision support system that generates recommendations on patient’s treatment and further studies that are mandatory according to regulations, based on patient’s electronic medical record. The functionality of MDSS is based on knowledge base (KB), which consists of clinical guidelines formalized as mathematical logic operators. The system is also able to control how medical staff fulfills generated diagnostic and treatment prescriptions: deadlines, duration and order of actions, mandatory studies from regulations, drug therapy, etc. The assessment of the system was conducted on more than 2000 CKD patients on chronic dialysis from 30 “Nefrosovet” hemodialysis clinics. Results MDSS was developed using process-based management principles and was successfully employed on representative sample of patients with CKD on chronic dialysis from “Nefrosovet” hemodialysis clinics. The system generated recommendations for this group of patients for the duration of 6 months. Recommendations were related to diagnostics and treatment of patients, specific diagnoses, scheduled and additional studies, drug prescriptions (where dose was set by a physician), etc. As a result of using a system, we were able to reduce average treatment cost of patient by 14% for anemia related treatment and by 36% for BMD related treatment. At the same time the treatment quality has improved from an average of 46% to 60% of patients in target ranges of key laboratory parameters after 6 months of system usage. Treatment cost has reduced due to the timely start and gradual cancelation (by indications) of drug therapy, which also allowed us to avoid overadministration of unnecessary drugs. Conclusion Deployment of developed MDSS for process-based management of patients with CKD on chronic dialysis with anemia and BMD complications has shown high cost efficiency. Furthermore, automated diagnostics, therapy and prescription fullfilment control is performed without loss of quality, sometimes even with the improvement of quality of treatment process.

Machine learning (ML) algorithms are data-driven and given a goal task and a prior experience dataset relevant to the task, one can attempt to solve the task using ML seeking to achieve high accuracy. There is usually a big gap in the understanding between an ML experts and the dataset providers due to limited expertise in cross disciplines. Narrowing down a suitable set of problems to solve using ML is possibly the most ambiguous yet important agenda for data providers to consider before initiating collaborations with ML experts. We proposed an ML-fueled pipeline to identify potential problems (i.e., the tasks) so data providers can, with ease, explore potential problem areas to investigate with ML. The autonomous pipeline integrates information theory and graph-based unsupervised learning paradigms in order to generate a ranked retrieval of top-k problems for the given dataset for a successful ML based collaboration. We conducted experiments on diverse real-world and well-known datasets, and from a supervised learning standpoint, the proposed pipeline achieved 72% top-5 task retrieval accuracy on an average, which surpasses the retrieval performance for the same paradigm using the popular exploratory data analysis tools. Detailed experiment results with our source codes are available at: https://github.com/jpastorino/heyml.

BackgroundThis study proposes a decision support system created in collaboration with machine learning experts and ophthalmologists for detecting keratoconus (KC) severity. The system employs an ensemble machine model and minimal corneal measurements.MethodsA clinical dataset is initially obtained from Pentacam corneal tomography imaging devices, which undergoes pre-processing and addresses imbalanced sampling through the application of an oversampling technique for minority classes. Subsequently, a combination of statistical methods, visual analysis, and expert input is employed to identify Pentacam indices most correlated with severity class labels. These selected features are then utilized to develop and validate three distinct machine learning models. The model exhibiting the most effective classification performance is integrated into a real-world web-based application and deployed on a web application server. This deployment facilitates evaluation of the proposed system, incorporating new data and considering relevant human factors related to the user experience.ResultsThe performance of the developed system is experimentally evaluated, and the results revealed an overall accuracy of 98.62%, precision of 98.70%, recall of 98.62%, F1-score of 98.66%, and F2-score of 98.64%. The application's deployment also demonstrated precise and smooth end-to-end functionality.ConclusionThe developed decision support system establishes a robust basis for subsequent assessment by ophthalmologists before potential deployment as a screening tool for keratoconus severity detection in a clinical setting.

Medical decision support systems (MDSS) are computer-based programs that analyse data within a patient's healthcare records to provide questions, prompts, or reminders to assist clinicians at the point of care. Inputting a patient's data, symptoms, or current treatment regimens into an MDSS, clinicians are assisted with the identification or elimination of the most likely potential medical causes, which can enable faster discovery of a set of appropriate diagnoses or treatment plans. Explainable AI (XAI) is a "white box" model of artificial intelligence in which the results of the solution can be understood by the users, who can see an estimate of the weighted importance of each feature on the model's predictions, and understand how the different features interact to arrive at a specific decision. This book discusses XAI-based analytics for patient-specific MDSS as well as related security and privacy issues associated with processing patient data. It provides insights into real-world scenarios of the deployment, application, management, and associated benefits of XAI in MDSS. The book outlines the frameworks for MDSS and explores the applicability, prospects, and legal implications of XAI for MDSS. Applications of XAI in MDSS such as XAI for robot-assisted surgeries, medical image segmentation, cancer diagnostics, and diabetes mellitus and heart disease prediction are explored.

Objective: to evaluate the possibility of using artificial neural networks for integration into the medical decision support system as an optimization of outpatient management of patients with COPD.Materials and methods: a dynamic followup of 150 patients with chronic obstructive pulmonary disease, registered at the dispensary for the underlying disease, who completed the outpatient stage of pulmonary rehabilitation after a moderate exacerbation was carried out. The material of the study was a universal questionnaire of 69 indicators, including anamnesis, clinic, laboratory and instrumental diagnostics. A four-layer neural network has been created: the first two layers — 69 neurons, the third layer — 34 neurons and the last layer — 3 neurons.Results: the software was used in the Java programming language using the Encog 3.4 module.Conclusion: the use of the capabilities of artificial neural networks for integration into the medical decision support system in the outpatient management of patients with chronic obstructive pulmonary disease has shown high specificity. The predictive model is implemented in the form of a computer program: "The program for predicting an unfavorable outcome, the development of cardiovascular complications and the effectiveness of rehabilitation measures in patients with chronic obstructive pulmonary disease (CardioRisk)" and was introduced into the work of outpatient polyclinic institutions in Rostov-on-Don.

This special issue is devoted to invited articles on machine learning (ML). Most of the articles included in this issue were also presented in a special session on ML at the 8th International Conference on Human-Compute r Interaction that was held at Yokohama, Japan in July 1995 (Anzai, Ogawa, & Mori, 1995). Since the publication of the first volume of Machine Learning: An Artificial Intelligence Approach (Michalski, Carbonell, & Mitchell, 1983), ML has progressed significantly and several applications have been reported, whereas several others have remained unpublished. In the same volume, the Nobel prize winner Herbert A. Simon places ML in context with learning by stating that learning denotes changes in the system that are adaptive in the sense that they enable the system to do the same task or tasks drawn from the same population more efficiently and more effectively the next time. Many scientific journals and international conferences have hosted special sections and sessions reporting on ML or on applications of ML. Knowledge acquisition, planning, scheduling, decision support, transportation, medicine, and engineering, among others, compose the domains in which ML has been both applied, proved effective, and continues to do so. An attempt to review all ML applications or theory developments would render this introduction or even the special issue endless. In part, a goal of this issue is to extend hands between the two communities: human-computer interaction (HCI) and ML. To a large degree, both share a common goal: Each one tries to improve the human performance and adaptability to changing conditions of some system. Enhancing systems with learning ability may prove conducive to building better systems. Humans come in life with built-in learning potential and excluding artifacts from learning may seriously impede user acceptability of new technology. The article by Moustakis, Lehto, and Salvendy captures expert judgment about a critical question: Which ML method should be used for a given task? The article is based on an extensive survey of ML experts and statistical analysis of responses. It also kicks off the special issue because it briefly introduces the reader to the various ML methods and tasks in which ML may be used. The article by Yoshida and Motoda presents a framework for using ML to automate user modeling and behavior in a user adaptive interface system. It uses

Machine and deep learning are used worldwide. Machine Learning (ML) and Deep Learning (DL) are playing an increasingly important role in the healthcare sector, particularly when combined with big data analytics. Some of the ways that ML and DL are being used in healthcare includePredictive Analytics,Medical Image Analysis,Drug Discovery,Personalized Medicine,and Electronic Health Records (EHR) Analysis.It has become one of the advanced and popular tool forcomputer sciencedomain.' Theadvancementof ML and DL forvariousfieldshas opened new avenues for research and development. It could revolutionize prediction and decision-makingcapabilities. Due to increased awarenessabout the ML and DL in the healthcare, it has become one of the vital approaches for thesector. High-volumeof unstructured, and complex medical imaging data from health monitoring devices, gadgets, sensors, etc. Is the biggest trouble forhealthcare sector. The current study uses analysis to examine research trends inadoption of machine learning and deep learningapproaches in the healthcare sector. The WoS database for SCI/SCI-E/ESCIjournals areused as the datasetsfor the comprehensive analysis. Apart from these various search strategy are utilised for therequisitescientificanalysis of the extracted research documents. Bibliometrics R statistical analysis is performed for year-wise, nation-wise, affiliation-wise, research area, sources, documents, and authorbased analysis. VOS viewer software is used to create author, source, country, institution, global cooperation, citation, co-citation, and trending term co-occurrence networks. ML and DL, combined with big data analytics, have the potential to revolutionize healthcare by improving patient outcomes, reducing costs, and accelerating the development of new treatments, so the currentstudy will help academics, researchers, decision-makers, and healthcare professionals understand and direct research.

Machine learning (ML) is being applied to a diverse and ever-growing set of domains. In many cases, domain experts - who often have no expertise in ML or data science - are asked to use ML predictions to make high-stakes decisions. Multiple ML usability challenges can appear as result, such as lack of user trust in the model, inability to reconcile human-ML disagreement, and ethical concerns about oversimplification of complex problems to a single algorithm output. In this paper, we investigate the ML usability challenges that present in the domain of child welfare screening through a series of collaborations with child welfare screeners. Following the iterative design process between the ML scientists, visualization researchers, and domain experts (child screeners), we first identified four key ML challenges and honed in on one promising explainable ML technique to address them (local factor contributions). Then we implemented and evaluated our visual analytics tool, Sibyl, to increase the interpretability and interactivity of local factor contributions. The effectiveness of our tool is demonstrated by two formal user studies with 12 non-expert participants and 13 expert participants respectively. Valuable feedback was collected, from which we composed a list of design implications as a useful guideline for researchers who aim to develop an interpretable and interactive visualization tool for ML prediction models deployed for child welfare screeners and other similar domain experts.

This presentation illustrates how big data forces change on algorithmic techniques and the goals of machine learning, bringing along challenges and opportunities. 1. The theoretical foundations of statistical machine learning traditionally assume that training data is scarce. If one assumes instead that data is abundant and that the bottleneck is the computation time, stochastic algorithms with poor optimization performance become very attractive learning algorithms. These algorithms quickly became the backbone of large-scale machine learning and are the object of very active research. 2. Increasing the training set size cannot improve average errors indefinitely. However this diminishing returns problem vanishes if we measure instead the diversity of conditions in which the trained system performs well. In other words, big data is not an opportunity to increase the average accuracy, but an opportunity to increase coverage. Machine learning research must broaden its statistical framework in order to embrace all the (changing) aspects of real big data problems. Transfer learning, causal inference, and deep learning are successful steps in this direction.

Machine learning (ML) is the fastest growing field in computer science, and health informatics is among the greatest challenges. The goal of ML is to develop algorithms which can learn and improve over time and can be used for predictions. Most ML researchers concentrate on automatic machine learning (aML), where great advances have been made, for example, in speech recognition, recommender systems, or autonomous vehicles. Automatic approaches greatly benefit from big data with many training sets. However, in the health domain, sometimes we are confronted with a small number of data sets or rare events, where aML-approaches suffer of insufficient training samples. Here interactive machine learning (iML) may be of help, having its roots in reinforcement learning, preference learning, and active learning. The term iML is not yet well used, so we define it as “algorithms that can interact with agents and can optimize their learning behavior through these interactions, where the agents can also be human.” This “human-in-the-loop” can be beneficial in solving computationally hard problems, e.g., subspace clustering, protein folding, or k-anonymization of health data, where human expertise can help to reduce an exponential search space through heuristic selection of samples. Therefore, what would otherwise be an NP-hard problem, reduces greatly in complexity through the input and the assistance of a human agent involved in the learning phase.