Chris Nugent’s research while affiliated with University of Ulster and other places

... LLMs can process and understand natural language, allowing the development of chatbots capable of engaging in meaningful conversations with users and fragile people [26]. By integrating LLMs with HAR systems, chatbots take advantage of the activity data collected to provide activity recognition [27], customised support, such as reminders, evaluation of mental status [28], loneliness [29], and even emotional support [30] that improves responsible use of social care [31]. For example, a chatbot could remind a user to take their medication according to their daily routine and detected activity, or offer encouragement and motivation if it senses that the user is struggling with a particular task [30]. ...

Reference:

Enhancing Smart Environments with Context-Aware Chatbots using Large Language Models

... The data utilized originate primarily from public sources such as the monkepox dataset provided by Our World in Data. These datasets may suffer from inconsistencies in data collection, reporting delays, or missing information, all of which can potentially affect the predictive accuracy of the model [13,22,52]. For example, in cases of sparse data, the model may struggle to adequately capture the characteristics of disease transmission, leading to increased prediction errors. ...

Reference:

From COVID-19 to monkeypox: a novel predictive model for emerging infectious diseases

... Seniors are more likely to wear a smartwatch as the watch sensor is already integrated into our lives, thus eliminating discomfort or resistance that might arise from other monitoring devices. Nonetheless, when collected in real-life settings, wearable sensor data has many challenging attributes: missing data is expected due to battery issues; non-wear periods can be prolonged if the user forgets to wear the smartwatch; and, we cannot get enough labeled data to recognize activities as it would burden the users to label them [12]. ...

Reference:

Daily Routine Recognition from Longitudinal, Real-Life Wearable Sensor Data for the Elderly

... To address this challenge, existing methods include data augmentation [18], model simplification [19] and Few-Shot Learning (FSL) [20][21][22][23][24][25]. In the past few years, many influential studies in FSL of optical target recognition offer us inspiration, which aim to enhance model generalization with only a few training samples given, including Model Agnostic Meta-Learning (MAML) [21], Prototypical Network (ProtoNet) [22], Deep Nearest Neighbor Neural Network (DN4) [23] and Relation Network (RN) [24]. ...

Reference:

Few-shot Human Motion Recognition through Multi-Aspect mmWave FMCW Radar Data

... 63 Rapid antigen-detecting tests give a rapid qualitative positive or negative, and they include immunochromatographic (ICR) tests, enzyme immunoassays (EIA), and optical immunoassays (OIA). [64][65][66][67] A 2018 study evaluated the radiological findings of RSV infection among 400 patients. The most common chest radiograph abnormalities were interstitial prominence, airspace opacity, and hyperinflation. ...

Reference:

Respiratory syncytial virus: A review of current basic and clinical knowledge

... Integrating UWB with nearby sensors, such as wearables, provides richer contextual information, improving activity recognition accuracy, particularly in multioccupant settings [19,18]. Furthermore, incorporating fuzzy logic helps to manage uncertainty and differentiate concurrent activities [20,21]. Advanced hardware and processing techniques further optimise data analysis for improved HAR accuracy [22]. ...

Reference:

Enhancing Smart Environments with Context-Aware Chatbots using Large Language Models

... As the first objective we aim to develop a methodology to create adherence computing platforms that allow healthy behaviors to be monitored in the homes of multiple patients, based on the successful ACTIVA human activity recognition approach [26], capable of recognizing activities by multiple users in the same environment by means of wearable and ambient devices. This new methodology will provide healthcare professionals and patients with valuable information about daily activities and lifestyle habits, making it easier to foster healthy behaviors and prevent disease. ...

Reference:

A New Horizon in Healthcare: An Innovative Methodology for Sensor-Based Adherence Platforms in Home Monitoring of Key Treatment Indicators

... Subsequently, several researchers have proposed enhanced control charts. For instance, Patel & Divecha [4] presented the modified exponentially weighted moving average (MEWMA) control chart to detect small shifts in process mean, and Khan et al. [5] improved upon this with a generalized form of MEWMA. Abbas et al. [6] combined the CUSUM and EWMA control charts, demonstrating that the mixed CUSUM-EWMA control chart performs better than either individual chart. ...

Reference:

Performance Evaluation of Extended EWMA Chart for AR Model with Exogenous Variables

... Because of this diversity, employing a single model is impractical for comprehensively learning all facets of the data. Therefore, the rationale behind the use of multimodal stacking ensembles stems from their success in integrating multiple information sources for complex decision making in various medical machine learning tasks 15,16 . ...

Reference:

A multimodal stacked ensemble model for cardiac output prediction utilizing cardiorespiratory interactions during general anesthesia

... In modern society, children's cognitive and behavioral problems have received increasing attention, especially the executive function and motor performance problems of children with attention deficit hyperactivity disorder (ADHD), which have become a hot topic in many research fields [1,2,3,4]. ADHD is a common neurodevelopmental disorder, mainly characterized by inattention, impulsive behavior, and hyperactivity. According to global statistics, the incidence of ADHD in children is about 5-7%, and a considerable number of children show severe cognitive and behavioral dysfunction. ...

Reference:

The effect of intelligent monitoring of physical exercise on executive function in children with ADHD