A5056350327- Machine Learning and Algorithms
- Machine Learning and ELM
- Time Series Analysis and Forecasting
- Domain Adaptation and Few-Shot Learning
- Cardiovascular Function and Risk Factors
- Complex Systems and Time Series Analysis
- Physical Activity and Health
- Sepsis Diagnosis and Treatment
- Heart Failure Treatment and Management
- Traffic Prediction and Management Techniques
- Body Composition Measurement Techniques
- Complex Network Analysis Techniques
- Human Mobility and Location-Based Analysis
- Data Management and Algorithms
- Data Stream Mining Techniques
- Data Visualization and Analytics
- Healthcare Operations and Scheduling Optimization
- Financial Risk and Volatility Modeling
- Advanced Clustering Algorithms Research
- Neural Networks and Applications
- Face and Expression Recognition
- Video Analysis and Summarization
- Emergency and Acute Care Studies
- Opinion Dynamics and Social Influence
- Artificial Intelligence in Healthcare
British Heart Foundation
2021-2022
King's College London
2019-2022
University of Oxford
2017-2020
University of Surrey
2011-2018
Institute of Biomedical Science
2018
Our world and our lives are changing in many ways. Communication, networking, computing technologies among the most influential enablers that shape today. Digital data connected worlds of physical objects, people, devices rapidly way we work, travel, socialize, interact with surroundings, they have a profound impact on different domains, such as healthcare, environmental monitoring, urban systems, control management applications, several other areas. Cities currently face an increasing...
An increasing number of cities are confronted with challenges resulting from the rapid urbanization and new demands that a rapidly growing digital economy imposes on current applications information systems. Smart city enable authorities to monitor, manage, provide plans for public resources infrastructures in environments, while offering citizens businesses develop use intelligent services cities. However, providing such smart gives rise several issues, as semantic heterogeneity...
This paper investigates the application of transductive transfer learning methods for action classification. The scenario is that off-line video annotation retrieval. We show if a classification system can analyze unlabeled test data in order to adapt its models, significant performance improvement be achieved. applied it tennis games train and videos different nature. Actions are described using HOG3D features we used method based on feature re-weighting novel translation scaling.
Classification methods traditionally work under the assumption that training and test sets are sampled from similar distributions (domains). However, when such deployed in practise, conditions which data is acquired do not exactly match those of set. In this paper, we exploit fact it often possible to gather unlabeled samples a test/target domain order improve model built source We propose Adaptive Transductive Transfer Machines, approach problem by combining four types adaptation: lower...
Symptom Cluster Research is a major topic in Cancer Science. In spite of the several statistical and clinical approaches this domain, there not consensus on which method performs better. Identifying generally accepted analytical important order to be able utilize process all available data. paper we report secondary analysis cancer symptom data, comparing performance five Machine Learning (ML) clustering algorithms doing so. Based how well they separate specific subsets measurements select...
Heart failure with preserved ejection fraction (HFpEF) is thought to be highly prevalent yet remains underdiagnosed. Evidence-based treatments are available that increase quality of life and decrease hospitalization. We sought develop a data-driven diagnostic model predict from electronic health records (EHR) the likelihood HFpEF among patients unexplained dyspnea left ventricular EF.
Cities have been a thriving place for citizens over the centuries due to their complex infrastructure. The emergence of Cyber-Physical-Social Systems (CPSS) and context-aware technologies boost growing interest in analysing, extracting eventually understanding city events which subsequently can be utilised leverage citizen observations cities. In this paper, we investigate feasibility using Twitter textual streams events. We propose hierarchical multi-view deep learning approach...
Abstract Background Delay in identifying deterioration hospitalised patients is associated with delayed admission to an intensive care unit (ICU) and poor outcomes. For the HAVEN project (HICF ref.: HICF-R9–524), we have developed a mathematical model that identifies real time facilitates intervention of ICU outreach team. This paper describes system has been designed implement model. We used innovative technologies such as Portable Format for Analytics (PFA) Open Services Gateway initiative...
The inter-departmental interactions and coordination of resources are two essential components for realising a smart city platform. In this study, we investigated citizens' role in enhancing facilitating the delivery services by merging three key aspects research field, namely Internet People, Things Web Data. To end, developed hybrid approach to extract meaningful information find physical-cyber-social similarity cities. specific data sources used study were Twitter, road traffic...
Machine learning has been used to accurately recognise physical activity patterns; however, classifiers for recognising targeted bone loading exercises have not developed. PURPOSE: The purpose of this study was determine the accuracy machine models classifying intensity necessary adaption in older adults. METHODS: Triaxial accelerometer data collected from forty-four participants (60-70 yrs) wearing a GCDC X16-1C on their hip during three aerobics classes consisting impact aerobic performed...
This work develops techniques for the sequential detection and location estimation of transient changes in volatility (standard deviation) time series data. In particular, we introduce a class change algorithms based on windowed filter. The first method detects by employing convex combination two such filters with differing window sizes, that adaptively updated weight parameter is then used as an indicator instantaneous power changes. Moreover, proposed adaptive filtering readily extended to...
ABSTRACT Aims Heart failure with preserved ejection fraction (HFpEF) is thought to be highly prevalent yet remains underdiagnosed. We sought develop a data-driven diagnostic model predict from electronic health records (EHR) the likelihood of HFpEF among patients unexplained dyspnea and left ventricular EF. Methods & Results The derivation cohort comprised echocardiography results. Structured unstructured data were extracted using an automated informatics pipeline. Patients...