A hybrid neural network model for noisy data regression

A hybrid neural network model, based on the fusion of fuzzy adaptive resonance theory (FA ART) and the general regression neural network (GRNN), is proposed in this paper. Both FA and the GRNN are incremental learning systems and are very fast in network training. The proposed hybrid model, denoted as GRNNFA, is able to retain these advantages and, at the same time, to reduce the computational requirements in calculating and storing information of the kernels. A clustering version of the GRNN is designed with data compression by FA for noise removal. An adaptive gradient-based kernel width optimization algorithm has also been devised. Convergence of the gradient descent algorithm can be accelerated by the geometric incremental growth of the updating factor. A series of experiments with four benchmark datasets have been conducted to assess and compare effectiveness of GRNNFA with other approaches. The GRNNFA model is also employed in a novel application task for predicting the evacuation time of patrons at typical karaoke centers in Hong Kong in the event of fire. The results positively demonstrate the applicability of GRNNFA in noisy data regression problems.

Towards a diversity justice management model : integrating organizational justice and diversity management

Purpose – Contemporary organizations are increasingly paying attention to incorporate diversity management practices into their systems in order to promote socially responsible actions and equitable employment outcomes for minority groups. The aim of this paper is to seek to address a major oversight in diversity management literature, the integration of organizational justice principles.

Design/methodology/approach – Drawing upon the existing literature on workforce diversity and organizational justice, the authors develop a model based on normative principles of organizational justice for justice-based diversity management processes and outcomes.

Findings – The paper proposes that effective diversity management results from a decision-making process that meets the normative principles of organizational justice (i.e. interactional, procedural and distributive justice). The diversity justice management model introduced in this article provides important theoretical and practical implications for establishing more moral and just workplaces.

Research limitations/implications – The authors have not tested the conceptual framework of the diversity justice management model, and recommend future research to take up the challenge. The payoff for doing so is to enable the establishment of socially responsible workplaces where individuals, regardless of their background, are given an equal opportunity to flourish in their assigned jobs.

Practical implications – The diversity justice management model introduced in this paper provides organizational justice (OJ)-based guidelines for managers to ensure that OJ can be objectively benchmarked and discussed amongst diversity stakeholders to continuously improve actual and perceived OJ outcomes.

Social implications – The social implication of this conceptual paper is reduction of workforce marginalization and establishment of socially responsible organizations whereby those marginalized (e.g. people with disabilities) can effectively work in their organizations.

Originality/value – This is the first attempt to establish a diveristy justice management model, which incorporates normative principles of organizational justice into diversity management processes and outcomes.

Parameter estimation of the BOLD fMRI model within a general particle filter framework

This work demonstrates a novel Bayesian learning approach for model based analysis of Functional Magnetic Resonance (fMRI) data. We use a physiologically inspired hemodynamic model and investigate a method to simultaneously infer the neural activity together with hidden state and the physiological parameter of the model. This joint estimation problem is still an open topic. In our work we use a Particle Filter accompanied with a kernel smoothing approach to address this problem within a general filtering framework. Simulation results show that the proposed method is a consistent approach and has a good potential to be enhanced for further fMRI data analysis.

Using Pareto principle plus statistic methodology in establishing a cost estimating model

Previous experience and research indicated that the Pareto Principle (80/20 Principle) has been widely used in many industries to achieve more with less. The study described in this paper concurs that this principle can be applied to improve the estimating accuracy and efficiency, especially in design development stage of projects. In fact, establishing an effective cost estimating model to improve accuracy and efficiency in design development stage has been a subject, which has attracted many research attentions over several decades. For over almost 40 years, research studies indicate that using the 80/20 Principle is one of the approaches. However, most of these studies were built by assumption, theoretical analysis or questionnaire survey. The objective of this research is to explore a logical and systematic method to establish a cost estimating model based on the Pareto Principle. This paper includes extensive literatures review on cost estimating accuracy and efficiency in the construction industry that points out the current gap of knowledge area and understanding of the topical. These reviews assist in developing the direction for the research and explore the potential methodology of using the Pareto Principle in the new cost estimating model. The findings of this paper suggest that combining the Pareto Principle with statistical analysis could be used as the technique to improve the accuracy and efficiency of current estimating methods in design development stage.

Contextually Learnt Detection of Unusual Motion-Based Behaviour in Crowded Public Spaces

In this paper we are interested in analyzing behaviour in crowded publicplaces at the level of holistic motion. Our aim is to learn, without user input, strong scene priors or labelled data, the scope of ‘‘normal behaviour’’ for a particular scene and thus alert to novelty in unseen footage. The first contribution is a low-level motion model based on what we term tracklet primitives, which are scenespecific elementary motions. We propose a clustering-based algorithm for tracklet estimation from local approximations to tracks of appearance features. This is followed by two methods for motion novelty inference from tracklet primitives: (a) an approach based on a non-hierarchial ensemble of Markov chains as a means of capturing behavioural characteristics at different scales, and (b) a more flexible alternative which exhibits a higher generalizing power by accounting for constraints introduced by intentionality and goal-oriented planning of human motion in a particular scene. Evaluated on a 2 h long video of a busy city marketplace, both algorithms are shown to be successful at inferring unusual behaviour, the latter model achieving better performance for novelties at a larger spatial scale.

Working with 'housing variance' to model urban growth futures with inner metropolitan Melbourne

Continued population growth in Melbourne over the past decade has led to the development of a range of strategies and policies by State and Local levels of government to set an agenda for a more sustainable form of urban development. As the Victorian State government moves towards the development of 'Plan Melbourne', a new metropolitan planning strategy currently being prepared to take Melbourne forward to 2050, the following paper addresses the issue of how new residential built form will impact on and be accommodated in existing Inner Melbourne activity centres. Working with the prospect of establishing a more compact city in order to meet an inner city target of 90,000 new dwellings (Inner Metropolitan Action Plan - IMAP Strategy 5), the paper presents a 'Housing Variance Model' based on household structure and dwelling type. As capacity is progressively altered through a range of built form permutations, the research attempts to assess the impact on the urban morphology of a case study of four Major Activity Centres in the municipality of Port Phillip.

PGSW-OS: a novel approach for resource management in a semantic web operating system based on a P2P grid architecture

A web operating system is an operating system that users can access from any hardware at any location. A peer-to-peer (P2P) grid uses P2P communication for resource management and communication between nodes in a grid and manages resources locally in each cluster, and this provides a proper architecture for a web operating system. Use of semantic technology in web operating systems is an emerging field that improves the management and discovery of resources and services. In this paper, we propose PGSW-OS (P2P grid semantic Web OS), a model based on a P2P grid architecture and semantic technology to improve resource management in a web operating system through resource discovery with the aid of semantic features. Our approach integrates distributed hash tables (DHTs) and semantic overlay networks to enable semantic-based resource management by advertising resources in the DHT based upon their annotations to enable semantic-based resource matchmaking. Our model includes ontologies and virtual organizations. Our technique decreases the computational complexity of searching in a web operating system environment. We perform a simulation study using the Gridsim simulator, and our experiments show that our model provides enhanced utilization of resources, better search expressiveness, scalability, and precision. © 2014 Springer Science+Business Media New York.

Analytical model to locate the fluidisation interface in a solid-gas vacuum fluidised bed

Vacuum fluidised beds have a distinct advantage of being operated with reduced mass consumption of the fluidising media. However, a low quality of fluidisation reduces the opportunity to utilise the bubbling regime in vacuum fluidised beds. Fluidisation maps are often used to depict the interface between the quiescent, bubbling and slugging regimes inside a fluidised bed. Such maps have been obtained by visual observations of the fluidisation interface in transparent fluidised beds. For beds which are visually inaccessible fluidisation maps are difficult to obtain. The present work therefore attempts to model the interface travel in a vacuum fluidised bed. The pressure gradient due to the bed weight has been determined to be a main contributor for fluidisation/defluidisation under vacuum. A simple analytical model based on the pressure gradient (PG model) is developed to predict the interface location in a vacuum fluidised bed. For a segregated bed, the Gibilaro-Rowe (GR) model is modified and used to predict the jetsam layer growth along with the fluidisation interface. The predictions are compared with the experimental data for minimally and highly segregated particles and it is seen that for non-segregated powders the predictions are quite accurate. Lack of sufficient knowledge of bubble characteristics, however, impeded accurate prediction of the jetsam growth especially at high flow rates. However, an approximate complete fluidisation interface is successfully predicted using the GR-PG model. © 2014 Elsevier B.V.

Towards designing an email classification system using multi-view based semi-supervised learning

The goal of email classification is to classify user emails into spam and legitimate ones. Many supervised learning algorithms have been invented in this domain to accomplish the task, and these algorithms require a large number of labeled training data. However, data labeling is a labor intensive task and requires in-depth domain knowledge. Thus, only a very small proportion of the data can be labeled in practice. This bottleneck greatly degrades the effectiveness of supervised email classification systems. In order to address this problem, in this work, we first identify some critical issues regarding supervised machine learning-based email classification. Then we propose an effective classification model based on multi-view disagreement-based semi-supervised learning. The motivation behind the attempt of using multi-view and semi-supervised learning is that multi-view can provide richer information for classification, which is often ignored by literature, and semi-supervised learning supplies with the capability of coping with labeled and unlabeled data. In the evaluation, we demonstrate that the multi-view data can improve the email classification than using a single view data, and that the proposed model working with our algorithm can achieve better performance as compared to the existing similar algorithms.

A framework for cloud-based healthcare services to monitor noncommunicable diseases patient

Monitoring patients who have noncommunicable diseases is a big challenge. These illnesses require a continuous monitoring that leads to high cost for patients' healthcare. Several solutions proposed reducing the impact of these diseases in terms of economic with respect to quality of services. One of the best solutions is mobile healthcare, where patients do not need to be hospitalized under supervision of caregivers. This paper presents a new hybrid framework based on mobile multimedia cloud that is scalable and efficient and provides cost-effective monitoring solution for noncommunicable disease patient. In order to validate the effectiveness of the framework, we also propose a novel evaluation model based on Analytical Hierarchy Process (AHP), which incorporates some criteria from multiple decision makers in the context of healthcare monitoring applications. Using the proposed evaluation model, we analyzed three possible frameworks (proposed hybrid framework, mobile, and multimedia frameworks) in terms of their applicability in the real healthcare environment.

DNA methylation and the social gradient of osteoporotic fracture: a conceptual model

INTRODUCTION: Although there is a documented social gradient for osteoporosis, the underlying mechanism(s) for that gradient remain unknown. We propose a conceptual model based upon the allostatic load theory, to suggest how DNA methylation (DNAm) might underpin the social gradient in osteoporosis and fracture. We hypothesise that social disadvantage is associated with priming of inflammatory pathways mediated by epigenetic modification that leads to an enhanced state of inflammatory reactivity and oxidative stress, and thus places socially disadvantaged individuals at greater risk of osteoporotic fracture. METHODS/RESULTS: Based on a review of the literature, we present a conceptual model in which social disadvantage increases stress throughout the lifespan, and engenders a proinflammatory epigenetic signature, leading to a heightened inflammatory state that increases risk for osteoporotic fracture in disadvantaged groups that are chronically stressed. CONCLUSIONS: Our model proposes that, in addition to the direct biological effects exerted on bone by factors such as physical activity and nutrition, the recognised socially patterned risk factors for osteoporosis also act via epigenetic-mediated dysregulation of inflammation. DNAm is a dynamic modulator of gene expression with considerable relevance to the field of osteoporosis. Elucidating the extent to which this epigenetic mechanism transduces the psycho-social environment to increase the risk of osteoporotic fracture may yield novel entry points for intervention that can be used to reduce individual and population-wide risks for osteoporotic fracture. Specifically, an epigenetic evidence-base may strengthen the importance of lifestyle modification and stress reduction programs, and help to reduce health inequities across social groups. MINI ABSTRACT: Our conceptual model proposes how DNA methylation might underpin the social gradient in osteoporotic fracture. We suggest that social disadvantage is associated with priming of inflammatory signalling pathways, which is mediated by epigenetic modifications, leading to a chronically heightened inflammatory state that places disadvantaged individuals at greater risk of osteoporosis.

A GARCH model for testing market efficiency

In this paper we propose a generalised autoregressive conditional heteroskedasticity (GARCH) model-based test for a unit root. The model allows for two endogenous structural breaks. We test for unit roots in 156 US stocks listed on the NYSE over the period 1980 to 2007. We find that the unit root null hypothesis is rejected in 40% of the stocks, and only in four out of the nine sectors the null is rejected for over 50% of stocks. We conclude with an economic significance analysis, showing that mostly stocks with mean reverting prices tend to outperform stocks with non-stationary prices.

Ocular exposure to UV-B in sunlight: the Melbourne Visual Impairment Project model

Quantification of ocular exposure to ultraviolet-B radiation (UV-B) has become an important public health issue, with reports that the ozone layer is being depleted worldwide. Ocular exposure to UV-B is determined by ambient UV-B levels, the duration of outdoor exposure, the proportion of ambient UV-B that reaches the eye, and the use of ocular protection. We have developed a simplified model for quantifying lifetime ocular UV-B exposure that can be used in large epidemiological surveys. Exposure to UV-B is assessed and quantified using a model based on personal exposure over the six summer months. Data available for a population-based sample of 1150 people in the age range 40-98 years revealed a distribution in average annual lifetime ocular UV-B exposure similar to that reported in a previous study on which this model is based, and also demonstrate that people can recall lifetime personal behaviour related to ocular protection. It takes 12 minutes on average to collect these data. This model can be employed by researchers worldwide for uniform assessment of ocular UV-B exposure.

Hybrid controller with the combination of FLC and neural network-based IMC for nonlinear processes

This work presents a hybrid controller based on the combination of fuzzy logic control (FLC) mechanism and internal model-based control (IMC). Neural network-based inverse and forward models are developed for IMC. After designing the FLC and IMC independently, they are combined in parallel to produce a single control signal. Mean averaging mechanism is used to combine the prediction of both controllers. Finally, performance of the proposed hybrid controller is studied for a nonlinear numerical plant model (NNPM). Simulation result shows the proposed hybrid controller outperforms both FLC and IMC.