Browsing by browse by HERDC Category Description "E1 Refereed Scholarly Conference Publication"

Many believe that data mining poses the biggest privacy challenge in the next decade. In this paper we concentrate on threats to privacy arising from the possibility of combining aggregate data to deduce the confidential individual values. We determine the maximum number of sum totals that can be disclosed without leading to a 2-compromise in a 1-dimensional database for range queries.

Many believe that data mining poses the biggest privacy challenge in the next decade. In this paper we concentrate on threats to privacy arising from the possibility of combining aggregate data to deduce the confidential individual values. We determine the maximum number of sum totals that can be disclosed without leading to a 2-compromise in a 1-dimensional database for range queries.

The Fifth Generation (5G) of mobile phone technology is gradually witnessing deployment in many parts of the world. New Zealand is also expected to start 5G deployment by the end of 2019. This paper reviews the progress made to date regarding the introduction of 5G specifically in New Zealand. This paper explores several technical and non-technical issues that relate to the 5G debate that is ongoing across the country. Of particular interest are topics that are associated with government's policy and actions towards 5G, and the perception of the general public about this upcoming technology.

Introduction: Multiple Myeloma (MM) is an incurable hematological malignancy affecting plasma cells marked by highly heterogeneous survival rate. Relapse is a significant impediment to the successful treatment of MM clinically. One of the main causes for relapse in MM is the development of multidrug resistance (MDR) to cancer chemotherapy. Currently, risk stratification to MM sub-groups and categorization of complete response to therapy are assessed based on molecular, cytogenetic markers using bone marrow biopsy as available systemic markers are incompetent in this regard. We are exploring the clinical significance of our recent in vitro and in vivo findings of a novel non-genetic basis to MDR whereby tiny vesicles called microparticles (MPs) shed from cancer cell's surface transfer MDR phenotype intercellularly. Microparticles isolated from the peripheral blood of patients who suffer from Multiple Myeloma will be phenotyped for resistance, adhesion and dissemination markers. Subsequently, these parameters will be correlated clinically to assess whether these characteristics are predictive of treatment outcome.

In this work, we proposed Software Defined Networking (SDN) based access control techniques for preventing unauthorised access to traffic flows in secure networks. We have developed an Access Control Application (ACA) for the SDN Controller for differentiating the flow requests from the user/devices that are classified at different security levels and configuring the routes with physical or virtual separation between the flows. This separation of flows makes it difficult for the malicious users with lower security clearance to access the flows that belong to the users with higher security clearance. Hence, our work significantly minimises the attack surface in secure environments. We will also discuss the prototype implementation of our model and some performance characteristics.

For multivariate, single-step genomic best linear unbiased prediction analyses fitting a breeding value model, it is often assumed that the proportions of total genetic variance accounted for by genomic markers and residual polygenic effects are the same for all traits. Different covariance matrices for the two types of genetic effects are readily taken into account by fitting them separately. However, this can lead to slow convergence rates in iterative solution schemes. We propose an alternative computing strategy which – exploiting a canonical transformation – allows for trait-specific covariances whilst directly fitting total genetic effects only. Its effects on convergence rates and gains in accuracy and bias of genomic evaluation compared to analyses assuming proportionality of covariance matrices are examined using a small simulation study. Results show comparatively little improvement in accuracies but worthwhile reductions in overdispersion of predicted genetic merits for genotyped individuals without phenotypes.

Rumen microbial profiles have been shown to be accurate predictors of methane emissions in a variety of species, however, it can be very costly and slow to generate a dataset with a sufficient number of individuals measured for methane who also have had rumen samples collected and processed into rumen microbial profiles for these benefits to be applied in industry. We evaluated the potential of combining datasets from New Zealand and Australian sheep to improve our ability to accurately predict methane emissions in Australian sheep. Prediction of Australian sheep methane emissions using rumen microbial profiles and phenotypes from New Zealand was possible, however, it was important to closely match the diets the sheep were fed to have confidence in the predictions. Prediction accuracies of Australian sheep methane emissions were higher when training on data collected on Australian sheep than training on New Zealand sheep; however augmentation of New Zealand data collected on a similar diet enabled more complex models to be run and an improvement in prediction accuracy.