OBJECTIVE: Clinical trial recruitment faces critical challenges with screen failure rates exceeding 80% in Alzheimer's disease (AD) trials. Traditional patient selection relies on expert consensus without systematic evaluation of trade-offs between statistical power, recruitment feasibility, safety, and cost. We developed a multi-objective optimization framework to systematically identify optimal eligibility criteria configurations that balance competing objectives in AD clinical trial design. METHODS: We implemented the Non-dominated Sorting Genetic Algorithm III (NSGA-III) to optimize patient selection criteria across three objectives: patient identification accuracy (F1 score), recruitment balance, and economic efficiency. The framework utilized National Alzheimer's Coordinating Center data comprising 2,743 participants with comprehensive clinical assessments and cerebrospinal fluid biomarker measurements. We optimized 14 eligibility parameters including age boundaries, cognitive thresholds, biomarker criteria, and comorbidity management policies. Statistical validation employed Monte Carlo simulation with 10,000 iterations, bootstrap analysis, and SHAP interpretability analysis. RESULTS: Optimization identified 11 Pareto-optimal solutions spanning F1 scores from 0.979 to 0.995 and eligible patient pools from 108 to 327. Compared to standard criteria selecting 101 participants, optimized approaches identified 102 participants with no significant demographic or clinical differences after multiple comparison correction. Monte Carlo simulation revealed mean cost savings of $1,048 per patient (95% CI: -$1,251 to $3,492), with 80.7% probability of positive savings but 19.3% risk of cost increases (SD = $1,208). Cross-validation demonstrated high precision (95.1%) with strategic selectivity (9.4% recall). SHAP analysis identified biomarker requirements as the dominant cost driver. Optimization algorithms converged toward solutions similar to expert-designed criteria, validating both computational and clinical approaches. CONCLUSION: Multi-objective optimization provides meaningful but incremental value through systematic validation and probabilistic efficiency enhancement rather than revolutionary transformation. The convergence toward established practice demonstrates that computational approaches serve as sophisticated validation tools that identify concrete yet uncertain efficiency improvements within existing frameworks. The substantial variability in projected outcomes establishes realistic expectations and highlights the importance of site-specific evaluation, particularly regarding recruitment infrastructure quality as the dominant determinant of success. This establishes a mature paradigm for evidence-based trial design optimization that enhances rather than replaces clinical expertise.

Cardiac patients with implantable cardioverter defibrillator devices frequently exhibit signs of anxiety and depression (termed "Distressed"). Early detection of these patients is vital for evaluation, intervention, and prevention against relapse. Considering the growing datasets relevant to distress, coupled with the evolution of machine learning methodologies, there exists a promising prospect to develop intelligent systems for the detection of distressed cardiac patients through written materials. In this context, data from two sources were collected: a questionnaire and text communication messages, acquired through the randomized ACQUIRE-ICD study of 168 participants. These textual messages were labelled as either Distressed or Non-Distressed based on questionnaire responses. Following preprocessing, the dataset facilitated the development of transformer-based classification models, including mBERT, XLM-RoBERTa, ÆLÆCTRA, and RøBERTa, as well as a hard voting ensemble method to classify patients into Distressed and Non-Distressed categories. To address imbalances in class distribution and dataset scarcity, a data augmentation method was employed. Results indicated the superior performance of the proposed hard voting ensemble, recording weighted metrics of 80% precision, 67% recall, 73% F1-score, and 75% accuracy. Notably, this ensemble correctly identified 67% of Distressed samples, while the most efficient base transformer, mBERT, identified 63% of Distressed samples.

BACKGROUND: Alcohol use disorder (AUD) causes significant morbidity, mortality, and injuries. According to reports, approximately 5% of all registered deaths in Denmark could be due to AUD. The problem is compounded by the late identification of patients with AUD, a situation that can cause enormous problems, from psychological to physical to economic problems. Many individuals suffering from AUD never undergo specialist treatment during their addiction due to obstacles such as taboo and the poor performance of current screening tools. Therefore, there is a lack of rapid intervention. This can be mitigated by the early detection of patients with AUD. A clinical decision support system (DSS) powered by machine learning (ML) methods can be used to diagnose patients' AUD status earlier. METHODS: This study proposes an effective AUD prediction model (AUDPM), which can be used in a DSS. The proposed model consists of four distinct components: (1) imputation to address missing values using the k-nearest neighbours approach, (2) recursive feature elimination with cross validation to select the most relevant subset of features, (3) a hybrid synthetic minority oversampling technique-edited nearest neighbour approach to remove noise and balance the distribution of the training data, and (4) an ML model for the early detection of patients with AUD. Two data sources, including a questionnaire and electronic health records of 2571 patients, were collected from Odense University Hospital in the Region of Southern Denmark for the AUD-Dataset. Then, the AUD-Dataset was used to build ML models. The results of different ML models, such as support vector machine, K-nearest neighbour, decision tree, random forest, and extreme gradient boosting, were compared. Finally, a combination of all these models in an ensemble learning approach was selected for the AUDPM. RESULTS: The results revealed that the proposed ensemble AUDPM outperformed other single models and our previous study results, achieving 0.96, 0.94, 0.95, and 0.97 precision, recall, F1-score, and accuracy, respectively. In addition, we designed and developed an AUD-DSS prototype. CONCLUSION: It was shown that our proposed AUDPM achieved high classification performance. In addition, we identified clinical factors related to the early detection of patients with AUD. The designed AUD-DSS is intended to be integrated into the existing Danish health care system to provide novel information to clinical staff if a patient shows signs of harmful alcohol use; in other words, it gives staff a good reason for having a conversation with patients for whom a conversation is relevant.

Alcohol Use Disorder (AUD) is a clinical diagnosis based on signs and symptoms that are related to excessive alcohol use and it increases the risk of many clinical conditions, psychological instabilities, and social issues. In this study, we aim to identify the comorbidities of Hazardous and Harmful drinkers. We obtained comorbidity networks for Hazardous, and Harmful drinkers using social network analysis techniques. Each network consists of several nodes that represent the diagnostic codes that patients were given during their hospitalization. To precisely identify the most exclusive comorbidities in each drinking group, we proposed a four-step process based on a machine learning algorithm and aggregation functions. Our findings show that the majority of the identified comorbidities in the Harmful drinking group are related to ICD-10 chapters XI, XIX, and XIII. The comorbidities of the Hazardous drinking group, however, did not present a similarly clear pattern.

Early detection of adverse events at hospitals could be useful in terms of reducing costs, morbidity, and mortality. Therefore, in this paper, we present a personalized real-time hybrid model based on Nonlinear Autoregressive Exogenous (NARX) model and Ensemble Learning (EL) to predict patients' severity during hospitalization at Emergency Departments (ED). This model utilizes vital signs of patients, including Pulse Rate (PR), Respiratory Rate (RR), Arterial Blood Oxygen Saturation (SpO2) and Systolic Blood Pressure (SBP), which are collected automatically during the treatment to predict the illness severity of hospitalized patients at ED in the next hour based on their vital signs of the previous two hours. Two EL algorithms, including Random Forest (RF) and Adaptive Boosting (AdaBoost) are considered to build hybrid models. The performance of NARX-EL models is compared with Auto Regressive Integrated Moving Average (ARIMA), combination of NARX and Linear Regression (LR), Support Vector Regression (SVR) and K-Nearest Neighbors Regression (KNN). The results show that our proposed hybrid models can predict patients' severity with significantly higher accuracy. It is also found that NARX-RF has the best performance in the prediction of sudden changes and unexpected adverse events in patients' vital signs (R2score =0.978, NRMSE =6.16%).

Prediction of alcohol use disorder (AUD) may reduce the number of deaths caused by alcohol-related diseases. However, prediction of AUD based on patients’ historical clinical data is still an open research objective. This study proposes a method to predict AUD from electronic health record (EHR) data through supervised machine learning. The study creates a dataset based on the combination of EHR data with patient reported data from 2,571 patients in the Region of Southern Denmark. After that, the dataset is labeled into two categories, AUD positive (457) and AUD negative (2,114). This unique dataset is used to validate the proposed method for prediction of AUD using machine learning methods based on historical clinical data from EHRs.

The prediction of Alcohol Use Disorder (AUD) may help to alleviate the number of deaths caused by alcohol related diseases, which had amounted to 3.3 million in 2014, worldwide. This article reports on the results of a scoping review of literature which focused on the prediction of AUD. A search in the academic databases including Medline, Web of Science and EBSCOhost had identified 28 articles which were published from 1980 to 2018, and which fulfilled our inclusion criteria related to the prediction of AUD. The findings suggest that research focusing on the prediction of AUD has been solid, with majority of the investigations focusing on genetics and family history, and psychological factors. It was observed that no study had tried to extract the predictor variables of AUD from their collected samples. Further, while a few studies had applied the machine learning approach in this domain, most investigations were based on statistical methods. Our review also suggests that compared to other regions where the rate of harmful drinking and mortality caused by alcoholism is high, Denmark is a country that has been less explored.

Chronic diseases such as heart and blood vessels are considered among the most common and serious reasons of mortality in the world. In Europe alone, over four million deaths a year (45% of all deaths) are caused by heart diseases [1]. In addition, chronic diseases are responsible for 70 % of United States deaths, and account for more than 75% of annual United States medical care cost [2]. For instance, Cardio Vascular Diseases (CVD) are considered the main cause for around 14.3% of total deaths in Denmark5, and it is the main cause for over 45% of the total death in Lebanon6. It is too costly to keep CVD patients under control locally within the vicinity of a healthcare unit. Thus, researchers recently started to realize the need for automated monitoring in health systems that are expected to reduce the overall death rate and cost associated with monitoring of patients. However, a general monitoring health system will not cover all diseases at once. Therefore, there is a pushing necessity for monitoring health systems which are dedicated to specific health cases. To contribute to the ongoing efforts, this work develops an automated system which could be customized for various chronic diseases. A mobile application based solution is proposed. Further, the work concentrates on CVD by conducting a survey in Lebanon to investigate the acceptance and awareness of ECG for remote monitoring of patients. The results are promising and reflect how specialists are aware of the need to utilize the rapid development in technology combined with the widespread usage of mobile phone which may be used as the main device to guarantee 24/7 communication link for ECG. Adopting ECG in the healthcare system will allow for capturing some valuable data which could guide the development of a recommendation system. This will issue necessary alerts to specialists and guidance to patients and their careers so that specialists could attend to the case on timely basis and patients with their careers could follow the recommendations to keep the case under control until the specialist becomes available. Finally, a secure forum based communication system will be developed to allow patients to share their experience and specialists to provide consultancy and guidance on demand.

Evaluation of network visualization tools in software engineering is a tricky task. There are a number of factors, contradictions, and preferences of investigative analysts that are to be kept in consideration while designing the experiment. Complexity in data, computational overhead to reach the targeted information and scarcity of a standard platform may slow down the investigation process. In this research paper, we have made the evaluation of some of the new features of our proposed framework, PEVNET, by conducting an experiment. There were twenty four participants who had evaluated the system. The experiment was performed in two phases. In the first phase, a usability evaluation and qualitative feedback was carried out to check whether the PEVNET framework provided adequate results to the users. The qualitative feedback was performed by considering two aspects: the ease of use and the functionality. We have conducted an evaluation of the newly inducted features into PEVNET. These include the 'Pie-chart feature', 'Trend analysis Feature', 'Graphical trend analysis feature', and 'Encircle feature'. In the second phase, the comparison of the PEVNET had been performed against some other state-of-the-art tools. These tasks were to be performed in the groups of participants. We found that the participants of the PEVNET group performed the tasks faster, in respect to the respective features, as compared to the other techniques used in the experiment. Further, we have found that the network visualization of the PEVNET framework, based on the experimental results, had gotten satisfactory feedback from the majority of the participants. The case study of Chicago Narcotics datasets was used. We believe that by evaluating the PEVNET in this research paper, we will be able to check the effectiveness of our proposed features.

No major criminal activity is possible without a comprehensive plot behind it. Detecting and understanding criminal activity has been a challenging task for the researchers in criminal networks. One important way of addressing those challenges has been visualization of criminal networks. We propose a framework called PEVNET in which existing visualization techniques for criminal networks are re-designed from a different perspective. Visualization features by way of merging, linking, and grouping of entity attributes is provided to criminal network investigators. Furthermore, we believe that the prevailing challenges to information visualization can be eliminated to a large extent by detecting evolving network patterns, which are extracted by way of visual analysis of criminal activity based on temporal data. Finally, the proposed framework will indicate the most central person in the network in a unique way, which will support the investigators' decision making.

SUMMARY In Model‐Driven Development (MDD), detection of model defects is necessary for correct model transformations. Formal verification tools and techniques can to some extent verify models. However, scalability is a serious issue in relation to verification of complex UML/OCL class diagrams. We have proposed a model slicing technique that slices the original model into submodels to address the scalability issue. A submodel can be detected as unsatisfiable if there are no valid values for one or more attributes of an object in the diagram or if the submodel provides inconsistent conditions on the number of objects of a given type. In this paper, we propose a novel feedback technique through model slicing that detects unsatisfiable submodels and their integrity constraints among the complex hierarchy of an entire UML/OCL class diagram. The software developers can therefore focus their revision efforts on the incorrect submodels while ignoring the rest of the model. Copyright © 2013 John Wiley & Sons, Ltd.

Model errors are a major concern in the paradigm of Model-Driven Development (MDD) because of model transformations and code generation. It is important to detect model errors before transformation as in the later stages it is harder to trace and fix such errors. Formal verification tools and techniques can check the correctness of a model, but their high computational complexity can limit their scalability. In this research, we present a tool named UMLtoCSP (UOST) that uses a UML/OCL Slicing Technique (UOST) to verify complex UML/OCL class diagram. The tool accepts UML class diagrams annotated with OCL constraints as input, breaks the original model m into m1, m2, m3,...,mn sub-models while abstracting unnecessary model elements.

In this paper, a new Cluster based Classification Model (CCM) for suspicious email detection and other text classification tasks, is presented. Comparative experiments of the proposed model against traditional classification models and the boosting algorithm are also discussed. Experimental results show that the CCM outperforms traditional classification models as well as the boosting algorithm for the task of suspicious email detection on terrorism domain email dataset and topic categorization on the Reuters-21578 and 20 Newsgroups datasets. The overall finding is that applying a cluster based approach to text classification tasks simplifies the model and at the same time increases the accuracy.

In this paper we present LanguageNet-a novel framework for processing unstructured text information from human generated content. The state of the art information processing frameworks have some shortcomings: modeled in generalized form, trained on fixed (limited) data sets, and leaving the specialization necessary for information consolidation to the end users. The proposed framework is the first major attempt to address these shortcomings. LanguageNet provides extended support of graphical methods contributing added value to the capabilities of information processing. We discuss the benefits of the framework and compare it with the available state of the art. We also describe how the framework improves the information gathering process and contribute towards building systems with better performance in the domain of Open Source Intelligence.

The abundance of information regarding any topic makes the Internet a very good resource. Even though searching the Internet is very easy, what remains difficult is to automate the process of information extraction from the available online information due to the lack of structure and the diversity in the sharing methods. Most of the times, information is stored in different proprietary formats, complying with different standards and protocols which makes tasks like data mining and information harvesting very difficult. In this paper, an information harvesting tool (heteroHarvest) is presented with objectives to address these problems by filtering the useful information and then normalizing the information in a singular non hypertext format. Finally we describe the results of experimental evaluation. The results are found promising with an overall error rate equal to 6.5% across heterogeneous formats.

This paper focuses on the problem of alias detection based on orthographic variations of Arabic names. Alias detection is the process to identify different variants of the same name. To detect aliases based on orthographic variations, the approximate string matching (ASM) algorithms are widely used that measure the similarities between two strings (i.e., the name and alias). ASM algorithms work well to detect various type of orthographic variations but still there is a need to develop techniques to detect correct aliases of Arabic names that occur due to the translation of Arabic names into English. An extension to widely used ASM algorithms is proposed to detect the name aliases that generate as a result of transliteration. This paper aims to improve the accuracy of the basic ASM algorithms in order to detect correct aliases. The experimental evaluation shows that proposed extension increases the accuracy of the basic algorithms to a considerable level.

A terrorist network is a special kind of social network with emphasis on both secrecy and efficiency. Such networks are intentionally structured to ensure efficient communication between members without being detected. A terrorist network can be modeled as a generalized network (graph) consisting of nodes and links. Techniques from social network analysis and graph theory can be used to identify key entities in the network, which is helpful for network destabilization purposes. Research on terrorist network analysis has mainly focuses on analysis of nodes, which is in contrast to the fact that the links between the nodes provide at least as much relevant information about the network as the nodes themselves. This paper presents a novel method to analyze the importance of links in terrorist networks inspired by research on transportation networks. The link importance measure is implemented in CrimeFighter Assistant and evaluated on known terrorist networks.

Model defects are a significant concern in the Model-Driven Development (MDD) paradigm, as model transformations and code generation may propagate errors to other notations where they are harder to detect and trace. Formal verification techniques can check the correctness of a model, but their high computational complexity can limit their scalability. In this paper, we consider a specific static model (UML class diagrams annotated with unrestricted OCL constraints) and a specific property to verify (satisfiability, i.e., "is it possible to create objects without violating any constraint?"). Current approaches to this problem have an exponential worst-case runtime. We propose a technique to improve their scalability by partitioning the original model into submodels (slices) which can be verified independently and where irrelevant information has been abstracted. The definition of the slicing procedure ensures that the property under verification is preserved after partitioning.

Open hypermedia systems hold significant potential to facilitate the development of collaborative design applications. In particular, it is the flexible approach utilized by these systems to provide infrastructure services, their ability to accommodate alternative hypermedia domains, and their facilities for information integration across application boundaries that give open hypermedia systems this potential. Current open hypermedia systems, however, have not realized their full potential to support collaborative design. They lack capabilities in one or more of these important areas. Improvements are especially needed in the collaborative infrastructure services provided by open hypermedia systems and in their support for alternative hypermedia domains. A prototype open hypermedia system is currently being extended to address these shortcomings of existing systems.

Many manufacturing enterprises are now trying to introduce various forms of flexible work organizations on the shop floor. However, existing computer-based production planning and control systems pose severe obstacles for autonomous working groups and other kinds of shop floor control to become reality. The research reported in this paper is predicated on the belief that the CSCW approach could offer a strategy for dealing with this problem. The paper describes the field work and its constructive outcome: a system that assists shop-floor teams in dealing with the complexities of day-to-day production planning by supporting intelligent and responsible workers in their situated coordination activities on the shop floor.

Development of hypermedia systems is a complex matter. The current trend toward open, extensible, and distributed multiuser hypermedia systems adds additional complexity to the development process. As a means of reducing this complexity, there has been an increasing interest in hyperbase management systems that allow hypermedia system developers to abstract from the intricacies and complexity of the hyperbase layer and fully attend to application and user interface issues. Design, development, and deployment experiences of a dynamic, open, and distributed multiuser hypermedia system development environment called Hyperform is presented. Hyperform is based on the concepts of extensibility, tailorability, and rapid prototyping of hypermedia system services. Open, extensible hyperbase management systems permit hypermedia system developers to tailor hypermedia functionality for specific applications and to serve as a platform for research. The Hyperform development environment is comprised of multiple instances of four component types: (1) a hyperbase management system server, (2) a tool integrator, (3) editors, and (4) participating tools. Hyperform has been deployed in Unix environments, and experiments have shown that Hyperform greatly reduces the effort required to provide customized hyperbase management system support for distributed multiuser hypermedia systems.

Software development environments are examples of complex computer applications characterized by heterogeneity; they are composed of diverse information repositories, user interfaces, services and tools. This paper presents an approach for providing hypermedia linking services as a means for integration in these heterogeneous settings. The overall goal of the HyperDisco project is to provide a hypermedia framework for flexible software system integration. The basic idea in this approach is to allow different tools to be integrated in the hypermedia framework at different tool-dependent levels. Instead of providing a single model of integration that all tools must adhere to, we allow each tool to have its own specialized model of integration and its own specialized protocol for accessing the hypermedia services. This paper describes the approach (focusing on integration and modeling aspects) and presents a prototype which supports it. Preliminary experience with the HyperDisco prototype and its relationship to other systems is described.