The increasing number of Jordanian Internet users should naturally be reflected in e-commerce conversions. However, this is not the case. Although social-media users in Jordan are becoming more engaged and involved in social-media transactions, e-commerce activities have not experienced a similar trend. This issue has been identified in the literature as the e-commerce awareness paradox, wherein customers are partially aware but are not engaged. This thesis investigates the lack of e-commerce adoption in Jordan, with special emphasis on exploring the role of awareness in e-commerce adoption. Using a mixed-method approach comprising Archival Research, Survey, Interviews, Focus group, and Narrative Inquiry Ethnography research, this thesis has emphasised that awareness should not be perceived as a holistic entity that influences the engagement of e-commerce, but rather as multiple degrees of awareness associated with different e-commerce processes. This thesis contributes to the Information Systems body of research by providing a new quantitative mapping technique. It projects a non-integral view of awareness on the ecommerce processes, which has resulted in identifying four distinctive awareness levels. Those four levels of awareness are Awareness of Products/Services (AOP/S), Awareness of Brand (AOB) Awareness of Payment (AOP), and Awareness of Delivery (AOD). Consequently, this mapping technique helped to put awareness into perspective as to what process of e-commerce needs to be tackled to appropriately help e-commerce practitioners to identify where they need to focus on the online acquisition journey. In addition, it provides an innovative level of stakeholder involvement that helped in identifying the role of the businesses and stakeholders in facilitating e-commerce. This has helped to provide several solutions to overcome many barriers that impediment the adoption of e-commerce in Jordan. Finally, this thesis provides a generic framework through the deployment of e-commerce processes, levels of awareness, and, stakeholders’ involvement which helps to integrate the process of successful e-commerce adoption within the Jordanian context. It contributes to gain a better understanding of the different level of awareness, which needs to be considered in each process of e-commerce for successful online shopping.

The work presented in this thesis focuses on exploring the potential of the use and development of mobile location-based services at outdoor cultural heritage sites. This PhD research investigated how people use mobile and wearable technologies for learning purposes with respect to cultural heritage sites. A user-centred design approach was adopted in this thesis using the socio-cognitive engineering methodology. Three empirical studies (field studies) were conducted with the aim of capturing users’ requirements adopting mixed methods. The studies were conducted sequentially using focus group, questionnaire and interview techniques; the focus group and questionnaire were conducted with potential end-users (learners), and the interviews were conducted with officials of cultural heritage and potential end-users. The studies with end-users were carried out to investigate their habits, behaviours and attitudes when using mobile and wearable technologies at outdoors cultural heritage sites. The official staff were interviewed to extract their opinions regarding using such services at their sites as well as find out what technologies they actually used to present information to their visitors. The results of the field studies led to the development of a theoretical framework, FoSLE, supported by the learning theories. FoSLE is introduced for designing smart and ubiquitous learning environments based on mobile and wearable technologies for outdoor cultural heritage sites. The framework was further analysed to pull out general requirements (GRs) (high-level requirements – more abstract) to be adopted in developing new technology supported artefacts. Four scenarios were developed based on the identified requirements to depict the context of use as well as to draw out a list of low-level requirements (LRs), i.e. detailed requirements. The LRs informed the design of a proof-of-concept, a smart and ubiquitous learning environment based on mobile and wearable technologies, SmartC. SmartC was evaluated in the field in two cycles using experts of human-computer interaction and potential end-users (learners). A combination of observation and interview techniques were used in the evaluation studies alongside the cognitive walkthrough method in the expert study and a usability questionnaire in the user study. The results of the evaluation studies revealed that SmartC is user-friendly and suitable for learning. The results of the evaluation studies contributed to the enhancement of the list of LRs, which consequently led to devise a list of design recommendations. The list of the design recommendations was designed to assist researchers and designers in designing and developing smart and ubiquitous learning environments based on mobile and wearable technologies. This PhD research introduces two main contributions to add to the academic knowledge, which are: 1. FoSLE: a theoretical framework for smart and ubiquitous learning environments utilising mobile location-based services and wearable computing. 2. A list of design recommendations for designing smart and ubiquitous learning environments utilising mobile location-based services and wearable computing.

Face recognition is a mature and reliable technology for identifying people. Due to high-definition cameras and supporting devices, it is considered the fastest and the least intrusive biometric recognition modality. Nevertheless, effective spoofing attempts on face recognition systems were found to be possible. As a result, various anti-spoofing algorithms were developed to counteract these attacks. They are commonly referred in the literature a liveness detection tests. In this research we highlight the effectiveness of some simple, direct spoofing attacks, and test one of the current robust liveness detection algorithms, i.e. the logistic regression based face liveness detection from a single image, proposed by the Tan et al. in 2010, against malicious attacks using processed imposter images. In particular, we study experimentally the effect of common image processing operations such as sharpening and smoothing, as well as corruption with salt and pepper noise, on the face liveness detection algorithm, and we find that it is especially vulnerable against spoofing attempts using processed imposter images. We design and present a new facial database, the Durham Face Database, which is the first, to the best of our knowledge, to have client, imposter as well as processed imposter images. Finally, we evaluate our claim on the effectiveness of proposed imposter image attacks using transfer learning on Convolutional Neural Networks. We verify that such attacks are more difficult to detect even when using high-end, expensive machine learning techniques.

No description available.

Over the last decade, networks have become increasingly advanced in terms of size, complexity and the level of heterogeneity, due to increase of number of users, devices and implementation of cloud among big enterprises and developing smart cities. As networks become more complicated, the existing client-server paradigm suffers from problems such as delay, jitter, bad quality of service, insufficient scalability, availability and flexibility. The appearance of mobile agents' technology is getting popular as means for an efficient way to access remote resources on computer networks. Mobile Agent- systems usually benefit from the following: asynchronous execution, dynamic adaptation, fault-tolerance improvement in network latency, protocol encapsulation, reduction in network load and robustness. However, one of the major technical obstacles to a wider acceptance of the mobile agent is security which is the modus operandi to protect the mobile agents against malicious hosts. This work proposes how the Mobile Agents (MA), supported by a new solid models (detection and protection), can present a new way of securing mobile agents against malicious hosts. The work contributes in proposing a new computing model for protection against malicious hosts. This model is based on trust, which is a combination of two kinds of trust: policy enforcement and control and punishment. The originality of this model is the introduction of the concept of setting up an active storage element in the agent space, called as "home away from home", for partial result storage and separation as well as digital signing of the destination of the mobile agent. An efficient flooding detection scheme is developed by integrating the sketch technique with the Divergence Measures (Hellinger Distance, Chi-Square and Power Divergences). This type of integration can be considered unique in comparison with existing solutions over a Mobile Agent network. The sketch data- structure summarizes the mobile agent's process of calls generating into a fixed set of data for developing a probability model. The Divergence Measures techniques, combined with a Mobile Agent traffic, efficiently identifies attacks, by monitoring the distance between current traffic distribution and the estimated distribution, based on history information. Compared to the previous detection system and existing works, the proposed techniques achieve the advantages of higher accuracy and flexibility, to deal with low intensity attacks and the ability to track the period of attack. Simulation results are presented to demonstrate the performance of the proposed detection model. This work achieves in outperforming the existing detection solutions by tuning the Divergence Measures. An evaluation of the scheme is done via the receiver-operating characteristic (ROC). The work achieves in outperforming the existing detection solutions by tuning the Power Divergence with a value of β=2.2. With this value of β, the detection scheme leads to a very attractive performance in terms of True Positive Rate (100%), False Positive Rate (3.8%) and is capable of detecting low intensity attacks. Moreover, the Power Divergence with β=2.2 presents a better detection accuracy of 98.1% in comparison with Hellinger Distance (60%) and Chi-square (80%). Since the scenarios in consideration in this work can be reasonably related to any type of network, the strength of the proposed model can alternatively be applied to any enterprise network.

Mortality prediction for hospitalized patients is an important problem. Over the past few decades, several severity scoring systems and machine learning models have been developed for predicting mortality in hospitals in general, and in intensive care units in particular. However, early mortality prediction in intensive care remains an open challenge. Most research has focused on Severity of Illness scores or Data Mining models designed for risk estimation at least 24 or 48 hours after intensive care admission. In this study, we aim to provide a model that can predict mortality from the patient's early hours of admission and to reach a performance that is better than existing methods. This research is conducted on the Multiparameter Intelligent Monitoring in Intensive Care database. An in-depth analysis of the database has been conducted. Problem assumptions and initial attribute selections have been defined. Relevant data has been preprocessed, extracted and converted for data mining analysis. The thesis starts by presenting two initial studies to compare the performance of the different approaches for handling mortality prediction: (1) A comparative study of Severity of Illness scores for ICU mortality prediction and (2) A time-series analysis for ICU mortality prediction using data mining classification models. The two studies have enabled the provision of a pioneer framework for early mortality prediction named 'EMPICU', which investigates thoroughly the prediction effectiveness of data mining classification models, after 6 hours of admission. The framework is tested for classification performance with different attribute selections and different classification models handling both missing values and class imbalance problems. The best performing model is the EMPICU-Random Forests with the 7 physiological vital signs in addition to age with excellent performance with Area Under the Receiver Operating Characteristic curve of 0.90. The EMPICU-Random Forests model at 6 hours of admission outperformed Severity of Illness scores at 24 hours after admission, which indicates that the proposed model predicts earlier with higher performance.

An ideal upper-extremity prosthesis is expected to simultaneously decode users' intentions and deliver artificial somatosensory feedback. Among all the feasible feedback modalities, electrotactile stimulation remains the most promising solution due to its advantages of light weight, little noise and low power consumption. This thesis further enhances the existing electrotactile feedback strategies by proposing a haptics model, developing a portable electrotactile stimulation (ETS) system and establishing a virtual hand rehabilitation platform for implementation and evaluation. Firstly, a Gaussian distribution based haptics model is proposed to characterise the human fingertip's biomechanics, including a prediction model to estimate the contact force according to the fingertip deformation and a probabilistic model to describe force uncertainty. Experiments results reveal the non-linearity, dispersion and individual difference of the fingertip's mechanical behaviour. Secondly, a potable 16-channel ETS system with a wireless mode for transmission is developed to provide electrotactile feedback for clinical use. The proposed ETS system can generate stable current output with programmable stimulation parameters, including amplitude, frequency and pulse width. The ETS output waveforms and stability were evaluated by capability tests. Thirdly, a virtual hand rehabilitation platform is established to investigate the effect of electrotactile feedback on user training of hand grasping tasks. The platform consists of a surface electromyography (sEMG) acquisition module, a virtual grasping environment, and an ETS module. Experiments were conducted to evaluate the impact of electrotactile feedback on a closed-loop grasping control in comparison with the visual feedback and no feedback. The quantitative results show that the integration of electrotactile feedback can both reduce the duration of rehabilitation and improve the virtual grasping success rate in comparison with the no feedback condition while possessing a better practicality over visual feedback. In summary, the proposed electrotactile feedback centred research is validated in facilitating the user training and improving the rehabilitation performance. Despite the initial motivation of this thesis driven by the upper-extremity prostheses, the verified success of electrical stimulation is not confined to the hand rehabilitation scenarios but potentially applicable to a wider spectrum of applications, such as biomedical engineering and virtual reality.

Bitcoin is a digital currency based on a peer-to-peer network to propagate and verify transactions. Bitcoin is gaining wider adoption than any previous crypto-currency and many well-known businesses have begun accepting bitcoins as means of financial payments. However, the mechanism of peers randomly choosing logical neighbors without any knowledge about the underlying physical topology can cause a delay overhead in information propagation which makes the system vulnerable to double spend attacks due to inconsistencies in the blockchain. Aiming at alleviating the propagation delay problem, this thesis evaluates the concept of network clustering in tackling the propagation delay problem in the Bitcoin network throughout introducing a proximity-aware extensions to the current Bitcoin protocol, named Locality Based Clustering (LBC), Ping Time Based Clustering (BCBPT), Super Node Based Clustering (BCBSN), and Master Node Based Clustering (MNBC). The ultimate purpose of the proposed protocols, that are based on how clusters are formulated and nodes define their membership, is to improve the information propagation delay in the Bitcoin network. The proximity of connectivity in the Bitcoin network is increased in the LBC and BCBPT protocol by grouping Bitcoin nodes based on different criteria, physical location in LBC protocol and link latencies between nodes in the BCBPT. In the BCBSN protocol, geographical connectivity increases as well as the number of hops between nodes decreases through assigning one node to be a cluster head that is responsible for maintaining the cluster. Whereas, MNBC incorporates master node technology and proximity-awareness into the existing Bitcoin protocol with the aim of creating fully connected clusters based on physical Internet proximity. We show, through simulations, that the proposed approaches define better clustering structures that optimize the transaction propagation delay over the Bitcoin protocol. However, MNBC is more effective at reducing the transaction propagation delay compared to the BCBPT, LBC, and BCBSN. On the other hand, this thesis evaluates the resistance of the Bitcoin network and the proposed approaches against the partitioning attack. Even though the Bitcoin network is more resistant against partition attacks than the proposed approaches, more resources need to be spent to split the network in the proposed approaches especially with a higher number of nodes. Finally, this thesis introduces a novel methodology to measure the transaction propagation delay in the real Bitcoin network with the aim of validating any model of the Bitcoin network. Transaction propagation measurements show that the transaction propagation time is significantly affected by the number of connected nodes and the network topology which is not proximity defined. In addition, large-scale measurements of the real Bitcoin network are performed in thesis with the aim of providing an opportunity to parameterise any model of the Bitcoin network accurately. Furthermore, this thesis presents a simulation model of the Bitcoin peer-to-peer network which is an event based simulation.

In real-life classification problems, prior information about the problem and expert knowledge about the domain are often used to obtain reliable and consistent solutions. This is especially true in fields where the data is ambiguous, such as text, in which the same words can be used in seemingly similar texts but have a different meaning. Many of the proposed approaches rely on the bag-of-words representation, which loses the information about the structure of the text. In this thesis, a literature review of related works in text classification is provided which includes an overview of text classification methods. In addition, detailed review of related works of two text classification domains; search engines and question answering systems. The core contribution is divided into three main parts. The first contribution is the Customizable Grammar Framework for user-intent text classification (CGF) which employs a formal grammar approach and exploits domain-related information in a new way to represent text as a series of syntactic categories forming syntactic patterns. In addition, the proposed framework has been applied to different domains which resulted in the second and third contribution. The second contribution is the Grammar-Based Framework for Query Classification (GQC) which helped in the improvement of query identification and classification. The third contribution is the Grammar-Based Framework for Question Categorization and Classification (GQCC) which helped in the enhancement of question identification and classification. In addition, using different machine learning algorithms the overall results show that the proposed approach outperforms previous ones in terms of classification performance for query and question classifications. Finally, comparison of the classification performance with the state-of-the-art approaches has been conducted, results validate that the proposed approach improves the classification accuracy and the identification of the different types of queries and questions.

The topic is that hospital capacity for patient beds runs short. We wish to predict when this will occur. An inter-disciplinary approach to this problem is taken incorporating a Management Science/Operational Research perspective. The subject is the Perinatal Network System, which is described, analysed and modelled. An illustrative Case Study is taken of an English local neonatal unit, where new-born babies are cared for. The focus is High dependency cots. Recommendations produced are subject to human factors and implementation difficulties. In this work, Systems Thinking facilitates an understanding of relationships; Enterprise Architecture helps embed the context and address complexity; while Clinical Medicine underpins decision-making for individual patients. Research outputs include the Conceptual Research Framework, a Quality Metric, a Cot Predictor Tool and a Markovian model Design, which can be adapted in the future. Furthermore there is the milieu or connective ‘glue’, to provide unity. The methodology or Enterprise Modelling helps address the issue by facilitating understanding of both overview and detail.