ABSTRACT As technological advancements continue to shape decision support systems, algorithmic tools are increasingly utilised in career-related contexts. This research investigates how terminology influences individuals’ acceptance of algorithmic decision aids in career decision-making. We introduce the concept of algorithm term aversion, examining whether users’ preferences differ depending on how algorithms are labelled. Across three studies (N = 459), we explored preferences for algorithmically driven agents in various contexts: job applications (Study 1), future career advice (Study 2), and career advancement (Study 3). Findings reveal a consistent aversion to the term “artificial intelligence” across all contexts and outcome measures. However, broader algorithm aversion did not consistently emerge, suggesting terminology plays a critical role in user acceptance. Understanding how users respond to algorithmic terminology can inform the design of more user-friendly decision support systems, thereby enhancing the integration of AI into sensitive decision-making domains, such as career decisions.

Human factors in the use and efficacy of decision support technologies for type 1 diabetes: evidence from a randomized controlled trial.

Banana diseases remarkably influence the worldwide production of bananas. Innumerable studies have focused on timely recognition, prediction, and management of banana plant diseases using various chemical, biological, socio-economic, and AI-based methods. The survey scrutinizes 184 articles accumulated from Scopus, Web of Science, and Google Scholar using defined keywords. These findings reveal the global distribution of the previous studies on plant disease detection, the evolution of ML techniques, and the most frequently studied diseases. The literature shows a swift progress towards machine learning, deep learning, remote sensing, and IoT systems for banana plant disease detection. However, numerous AI models lack real-world validation, datasets are fragmented, and severity quantification mechanisms are understudied. The synthesis analyzes the strong dominance of CNN-based models, which account for the highest proportion of published works and remain the foundational architecture for banana disease detection. Countries such as India, China, the Philippines, Ecuador, and Indonesia have contributed significantly to disease detection. Despite notable progress, many existing systems still rely on single-source and limited datasets, which leads to a lack of cross-source robustness. Evolution of a robust framework integrating multiple datasets, explainable AI, decision support systems and socio-economic insights can lead to more enhanced farmer-friendly banana plant disease management in future This survey provides a detailed overview of the global research studies, highlighting key research gaps that need to be addressed and outlines future directions for building more reliable, interpretable, and comprehensive decision-support pipelines, which will guide the future research work.

Rendering knowledge graphs from aerospace dentistry processes for clinical decision support systems.

Early detection of Parkinson’s disease (PD) is essential for timely medical intervention and improving patient outcomes. Speech signal analysis offers a non-invasive, cost-effective, and easily deployable diagnostic pathway. However, achieving reliable early prediction remains challenging due to data imbalance, redundant features, and model instability. This study aims to develop an optimized and robust machine learning framework that enhances the predictive accuracy and stability of PD detection from speech data. An optimized machine learning model based on eXtreme Gradient Boosting (XGBoost) was developed for early PD prediction. The model’s hyperparameters were tuned using the Tree-structured Parzen Estimator (TPE), while Mutual Information (MI) was employed to select the most informative features from the speech dataset. To address class imbalance, the Adaptive Synthetic Sampling Approach for Imbalanced Learning (ADASYN) was applied to generate synthetic minority samples. Model performance and stability were evaluated using ten independent runs of Stratified 10-Fold Cross-Validation (SCV). The proposed framework achieved superior predictive performance with an average accuracy of 97.27%, precision of 98.79%, F1-score of 97.18%, recall of 95.77%, and ROC-AUC of 98.11% across multiple evaluations. Comparative analysis with similar studies demonstrated improved robustness, reliability, and balance between sensitivity and specificity. The integration of MI-based feature selection and ADASYN-based data augmentation significantly enhanced the performance and stability of the XGBoost model for early PD prediction. The proposed model demonstrates strong potential for clinical use as a decision support system, providing a low-cost, non-invasive, and remotely deployable tool for early PD diagnosis using patient speech signals.

Clinical and MRI variables in decision support systems for prostate MRI: A systematic review of decision support tools, nomograms, and risk models.

Building trustworthy large language model-driven generative recommender system for healthcare decision support: A scoping review of corpus sources, customization techniques, and evaluation frameworks.

Integrating rule-based NLP and large language models for statin information extraction from clinical notes.

Optimizing multimodal models for medical visual question answering: A comparative study of LoRA and AdaLoRA on VQA-RAD and SLAKE-VQA.

Identifying eligible patients for the Australian national lung cancer screening program in primary care: A cross-sectional study using clinical decision support systems and evaluating PLCOm2012 data quality.

A multi-granularity probabilistic linguistic decision support system with incomplete weight information affected by psychological disparity for smart water engineering

Revolutionizing pediatric obesity intervention strategies: From traditional growth reference tools to AI-enabled pediatric obesity clinical decision support systems.

A decision support system linking forest succession, habitat suitability and population dynamics of woodland caribou in the boreal forest of northern Ontario, Canada

Enhancing healthcare worker mental health via artificial intelligence-driven work process improvements: a scoping review.

Sensory integration (SI) typically follows a normal sequence of development. Its evaluation is crucial for evidence-based interventions. To examine the relationship between age and performance on the Evaluation in Ayres Sensory Integration (EASI) vestibular and proprioceptive tests and to compare linear and nonlinear models. Normative data from the EASI tests were collected and used for model-building to characterize age effects on SI abilities. Laboratory in Taiwan. Children (N = 209) ages 3-12 yr. Occupational therapists (N = 23) from various regions of Taiwan collected data from typically developing children. Linear and nonlinear (quadratic and hyperbolic) models were used to examine the relationship between age and test performance. Linear models accounted for 36%-43% of the variance in vestibular and proprioceptive test performance. Nonlinear models performed slightly better, explaining 42%-48% of the variance. The sum of squared errors was lower for nonlinear models (2,943-3,832) than for linear models (3,944-4,223). The hyperbolic model generally showed the highest R2 (explained variance) and the lowest root-mean-square error of approximation, except for the Joint Position test, where it did not outperform the other models. We developed a clinically applicable system for assessing vestibular and proprioceptive functions. The results showed marked developmental improvements, especially between ages 7 and 9 yr. The findings suggest that occupational therapists can use a hyperbolic model with EASI vestibular and proprioceptive tests to better identify SI challenges, which supports the development of evidence-based intervention plans. Plain-Language Summary: Sensory integration plays a vital role in children's typical development and tends to follow a predictable pattern as they grow. The Evaluation in Ayres Sensory Integration (EASI) is a tool that is used to assess children's sensory, motor, and motor planning skills. This study focused specifically on the vestibular and proprioceptive tests. The vestibular system in the inner ear maintains balance, posture, head position, and keeps our eyes stable during movement. Proprioception lets us sense limb position without seeing them. We assessed 209 children between the ages 3 and 12 years to explore how performance on these tests changes with age. We compared two types of models: (1) a linear model (which assumes steady, constant growth) and (2) a nonlinear model (which allows for more complex growth patterns). The results showed that the nonlinear hyperbolic model better captured how children's vestibular and proprioceptive abilities develop over time. One key finding was that these skills improved most noticeably between the ages of 7 and 9 years. We developed a clinical decision support system using the EASI vestibular and proprioceptive tests. For occupational therapists, this means that applying a hyperbolic model to interpret test results may lead to more accurate identification of sensory integration difficulties. This enhanced approach can guide the development of evidence-based intervention plans.

There is a growing proliferation of artificial intelligence (AI) in most spheres and sectors of contemporary society including social work. This article reports a survey of South Asian social workers' views on AI and social work including application domains, usefulness, risks and challenges, training needs, and future of the profession. The majority of respondents have suggested that social workers need training on machine learning, reinforcement learning, and natural language processing. A high proportion proposed that AI will redefine the profession's future through multisource data synthesis on client lifeworld contexts, analysis of macro- and organizational-level data for intervention, multiple domains of practical use, and AI-powered decision support systems to recommend interventions. They recommended having digital ethics committees and diverse stakeholder groups to review AI protocols and suggest modifications in case of algorithmic bias. They also highlighted the need for training sessions on the use of AI to ensure its responsible use in social work practice.

Measuring and predicting where and when pathologists focus their visual attention while grading whole slide images of cancer.

Automated Machine Learning in medical research: A systematic literature mapping study.

In Japan, prescription drug labeling has transitioned to a new structured format aimed at improving clarity and consistency, with full implementation in March 2024. Abnormal kidney function, a critical determinant of drug safety, necessitates clear and consistent contraindication labeling. However, current labeling practices have not been comprehensively evaluated. To systematically assess how kidney-related contraindications are described in Japanese package inserts under the new labeling format. We reviewed all electronically available prescription drug package inserts as of October 1, 2024. Using 20 kidney-related keywords, we extracted and analyzed statements in Sect.2 ("Contraindications"), categorizing them into four domains: type of impairment, severity, disease progression, and quantitative criteria. Additionally, a network diagram of co-occurring terms was developed to illustrate the consistency and diversity of terminology. A total of 233 kidney-related contraindication statements were identified across 182 pharmaceutical ingredients. Type of impairment was mentioned in 81.5%, severity in 54.9%, and quantitative criteria in 38.2%. However, the terminology and threshold values used were inconsistent. Terms such as "severe abnormal kidney function" were used without standardized definitions, and quantitative parameters (e.g., creatinine clearance, estimated glomerular filtration rate) varied across products. Despite regulatory efforts to enhance labeling structure, kidney-related contraindication descriptions in Japan remain variable and lack standardization. These inconsistencies may hinder safe prescribing practices and the integration of labeling into clinical decision support systems. Adoption of internationally harmonized terminology, such as KDIGO staging, and clearer regulatory guidance may improve the clinical utility of drug labeling.

MUTE-DSS: A digital-twin-based decision support system for minimizing underwater radiated noise in ship voyage planning