Machine learning techniques to improve software quality

A significant proportion of the cost of software development is due to software testing and maintenance. This is in part the result of the inevitable imperfections due to human error, lack of quality during the design and coding of software, and the increasing need to reduce faults to improve customer satisfaction in a competitive marketplace. Given the cost and importance of removing errors improvements in fault detection and removal can be of significant benefit. The earlier in the development process faults can be found, the less it costs to correct them and the less likely other faults are to develop. This research aims to make the testing process more efficient and effective by identifying those software modules most likely to contain faults, allowing testing efforts to be carefully targeted. This is done with the use of machine learning algorithms which use examples of fault prone and not fault prone modules to develop predictive models of quality. In order to learn the numerical mapping between module and classification, a module is represented in terms of software metrics. A difficulty in this sort of problem is sourcing software engineering data of adequate quality. In this work, data is obtained from two sources, the NASA Metrics Data Program, and the open source Eclipse project. Feature selection before learning is applied, and in this area a number of different feature selection methods are applied to find which work best. Two machine learning algorithms are applied to the data - Naive Bayes and the Support Vector Machine - and predictive results are compared to those of previous efforts and found to be superior on selected data sets and comparable on others. In addition, a new classification method is proposed, Rank Sum, in which a ranking abstraction is laid over bin densities for each class, and a classification is determined based on the sum of ranks over features. A novel extension of this method is also described based on an observed polarising of points by class when rank sum is applied to training data to convert it into 2D rank sum space. SVM is applied to this transformed data to produce models the parameters of which can be set according to trade-off curves to obtain a particular performance trade-off.

Overcoming the complexity of diagnostic problems due to sensor network architecture

In fault detection and diagnostics, limitations coming from the sensor network architecture are one of the main challenges in evaluating a system’s health status. Usually the design of the sensor network architecture is not solely based on diagnostic purposes, other factors like controls, financial constraints, and practical limitations are also involved. As a result, it quite common to have one sensor (or one set of sensors) monitoring the behaviour of two or more components. This can significantly extend the complexity of diagnostic problems. In this paper a systematic approach is presented to deal with such complexities. It is shown how the problem can be formulated as a Bayesian network based diagnostic mechanism with latent variables. The developed approach is also applied to the problem of fault diagnosis in HVAC systems, an application area with considerable modeling and measurement constraints.

Molecular characterisation of environmental enterococci derived from water samples and assessment of associated public health hazards

Enterococci are versatile Gram-positive bacteria that can survive under extreme conditions. Most enterococci are non-virulent and found in the gastrointestinal tract of humans and animals. Other strains are opportunistic pathogens that contribute to a large number of nosocomial infections globally. Epidemiological studies demonstrated a direct relationship between the density of enterococci in surface waters and the risk of swimmer-associated gastroenteritis. The distribution of infectious enterococcal strains from the hospital environment or other sources to environmental water bodies through sewage discharge or other means, could increase the prevalence of these strains in the human population. Environmental water quality studies may benefit from focusing on a subset of Enterococcus spp. that are consistently associated with sources of faecal pollution such as domestic sewage, rather than testing for the entire genus. E. faecalis and E. faecium are potentially good focal species for such studies, as they have been consistently identified as the dominant Enterococcus spp. in human faeces and sewage. On the other hand enterococcal infections are predominantly caused by E. faecalis and E. faecium. The characterisation of E. faecalis and E. faecium is important in studying their population structures, particularly in environmental samples. In developing and implementing rapid, robust molecular genotyping techniques, it is possible to more accurately establish the relationship between human and environmental enterococci. Of particular importance, is to determine the distribution of high risk enterococcal clonal complexes, such as E. faecium clonal complex 17 and E. faecalis clonal complexes 2 and 9 in recreational waters. These clonal complexes are recognized as particularly pathogenic enterococcal genotypes that cause severe disease in humans globally. The Pimpama-Coomera watershed is located in South East Queensland, Australia and was investigated in this study mainly because it is used intensively for agriculture and recreational purposes and has a strong anthropogenic impact. The primary aim of this study was to develop novel, universally applicable, robust, rapid and cost effective genotyping methods which are likely to yield more definitive results for the routine monitoring of E. faecalis and E. faecium, particularly in environmental water sources. To fullfill this aim, new genotyping methods were developed based on the interrogation of highly informative single nucleotide polymorphisms (SNPs) located in housekeeping genes of both E. faecalis and E. faecium. SNP genotyping was successfully applied in field investigations of the Coomera watershed, South-East Queensland, Australia. E. faecalis and E. faecium isolates were grouped into 29 and 23 SNP profiles respectively. This study showed the high longitudinal diversity of E. faecalis and E. faecium over a period of two years, and both human-related and human-specific SNP profiles were identified. Furthermore, 4.25% of E. faecium strains isolated from water was found to correspond to the important clonal complex-17 (CC17). Strains that belong to CC17 cause the majority of hospital outbreaks and clinical infections globally. Of the six sampling sites of the Coomera River, Paradise Point had the highest number of human-related and human-specific E. faecalis and E. faecium SNP profiles. The secondary aim of this study was to determine the antibiotic-resistance profiles and virulence traits associated with environmental E. faecalis and E. faecium isolates compared to human pathogenic E. faecalis and E. faecium isolates. This was performed to predict the potential health risks associated with coming into contact with these strains in the Coomera watershed. In general, clinical isolates were found to be more resistant to all the antibiotics tested compared to water isolates and they harbored more virulence traits. Multi-drug resistance was more prevalent in clinical isolates (71.18% of E. faecalis and 70.3 % of E. faecium) compared to water isolates (only 5.66 % E. faecium). However, tetracycline, gentamicin, ciprofloxacin and ampicillin resistance was observed in water isolates. The virulence gene esp was the most prevalent virulence determinant observed in clinical isolates (67.79% of E. faecalis and 70.37 % of E. faecium), and this gene has been described as a human-specific marker used for microbial source tracking (MST). The presence of esp in water isolates (16.36% of E. faecalis and 19.14% of E. faecium) could be indicative of human faecal contamination in these waterways. Finally, in order to compare overall gene expression between environmental and clinical strains of E. faecalis, a comparative gene hybridization study was performed. The results of this investigation clearly demonstrated the up-regulation of genes associated with pathogenicity in E. faecalis isolated from water. The expression study was performed at physiological temperatures relative to ambient temperatures. The up-regulation of virulence genes demonstrates that environmental strains of E. faecalis can pose an increased health risk which can lead to serious disease, particularly if these strains belong to the virulent CC17 group. The genotyping techniques developed in this study not only provide a rapid, robust and highly discriminatory tool to characterize E. faecalis and E. faecium, but also enables the efficient identification of virulent enterococci that are distributed in environmental water sources.

Large scale monocular vision-only mapping from a fixed-wing sUAS

This paper presents the application of a monocular visual SLAMon a fixed-wing small Unmanned Aerial System (sUAS) capable of simultaneous estimation of aircraft pose and scene structure. We demonstrate the robustness of unconstrained vision alone in producing reliable pose estimates of a sUAS, at altitude. It is ultimately capable of online state estimation feedback for aircraft control and next-best-view estimation for complete map coverage without the use of additional sensors.We explore some of the challenges of visual SLAM from a sUAS including dealing with planar structure, distant scenes and noisy observations. The developed techniques are applied on vision data gathered from a fast-moving fixed-wing radio control aircraft flown over a 1×1km rural area at an altitude of 20-100m.We present both raw Structure from Motion results and a SLAM solution that includes FAB-MAP based loop-closures and graph-optimised pose. Timing information is also presented to demonstrate near online capabilities. We compare the accuracy of the 6-DOF pose estimates to an off-the-shelfGPS aided INS over a 1.7kmtrajectory.We also present output 3D reconstructions of the observed scene structure and texture that demonstrates future applications in autonomous monitoring and surveying.

A portable approach for the surveillance of dengue virus-infected mosquitoes

Dengue virus is the most significant human viral pathogen spread by the bite of an infected mosquito. With no vaccine or antiviral therapy currently available, disease prevention relies largely on surveillance and mosquito control. Preventing the onset of dengue outbreaks and effective vector management would be considerably enhanced through surveillance of dengue virus prevalence in natural mosquito populations. However, current approaches to the identification of virus in field-caught mosquitoes require relatively slow and labor intensive techniques such as virus isolation or RT-PCR involving specialized facilities and personnel. A rapid and portable method for detecting dengue virus-infected mosquitoes is described. Using a hand held battery operated homogenizer and a dengue diagnostic rapid strip the viral protein NS1 was detected as a marker of dengue virus infection. This method could be performed in less than 30 min in the field, requiring no downstream processing, and is able to detect a single infected mosquito in a pool of at least 50 uninfected mosquitoes. The method described in this study allows rapid, real-time monitoring of dengue virus presence in mosquito populations and could be a useful addition to effective monitoring and vector control responses.

Traffic and climate change impacts on water quality : measuring build-up and wash-off of heavy metals and petroleum hydrocarbons

Understanding the impacts of traffic and climate change on water quality helps decision makers to develop better policy and plans for dealing with unsustainable urban and transport development. This chapter presents detailed methodologies developed for sample collection and testing for heavy metals and total petroleum hydrocarbons, as part of a research study to investigate the impacts of climate change and changes to urban traffic characteristics on pollutant build-up and wash-off from urban road surfaces. Cadmium, chromium, nickel, copper, lead, iron, aluminium, manganese and zinc were the target heavy metals, and selected gasoline and diesel range organics were the target total petroleum hydrocarbons for this study. The study sites were selected to encompass the urban traffic characteristics of the Gold Coast region, Australia. An improved sample collection method referred to as ‘the wet and dry vacuum system’ for the pollutant build-up, and an effective wash-off plan to incorporate predicted changes to rainfall characteristics due to climate change, were implemented. The novel approach to sample collection for pollutant build-up helped to maintain the integrity of collection efficiency. The wash-off plan helped to incorporate the predicted impacts of climate change in the Gold Coast region. The robust experimental methods developed will help in field sample collection and chemical testing of different stormwater pollutants in build-up and wash-off.

Spatial and temporal foraging patterns of Queensland fruit fly, Bactrocera tryoni(Froggatt) (Diptera: Tephritidae), for protein and implications for management

Fruit flies require protein for reproductive development and actively feed upon protein sources in the field. Liquid protein baits mixed with insecticide are used routinely to manage pest fruit flies, such as Bactrocera tryoni (Froggatt). However, there are still some gaps in the underpinning science required to improve the efficacy of bait spray technology. The spatial and temporal foraging behaviour of B. tryoni in response to protein was investigated in the field. A series of linked trials using either wild flies in the open field or laboratory-reared flies in field cages and a netted orchard were undertaken using nectarines and guavas. Key questions investigated were the fly's response to protein relative to: height of protein within the canopy, fruiting status of the tree, time of day, season and size of the experimental arena. Canopy height had a significant response on B. tryoni foraging, with more flies foraging on protein in the mid to upper canopy. Fruiting status also had a significant effect on foraging, with most flies responding to protein when applied to fruiting hosts. B. tryoni demonstrated a repeatable diurnal response pattern to protein, with the peak response being between 12:00–16:00 h. Season showed significant but unpredictable effects on fruit fly response to protein in the subtropical environment where the work was undertaken. Relative humidity, but not temperature or rainfall, was positively correlated with protein response. The number of B. tryoni responding to protein decreased dramatically as the spatial scale increased from field cage through to the open field. Based on these results, it is recommend that, to be most effective, protein bait sprays should be applied to the mid to upper canopies of fruiting hosts. Overall, the results show that the protein used, an industry standard, has very low attractancy to B. tryoni and that further work is urgently needed to develop more volatile protein baits.

Desalination by a solar thermal membrane distillation process

A solar thermal membrane distillation pilot plant was operated for over 70 days in field conditions. The pilot plant incorporated a single spiral wound permeate gap membrane distillation style of module. All energy used to operate the unit was supplied by solar hot water collectors and photovoltaic panels. The process was able to produce a distillate stream of product water with a conductivity less than 10 µS/cm. Feed water concentration varied from 2,400 µS/cm to 106,000 µS/cm. The process is expected to find application in the production of drinking water for remote island and arid regions without the consumption of electrical energy.

Review : alternative splicing (AS) of genes as an approach for generating protein complexity

Prior to the completion of the human genome project, the human genome was thought to have a greater number of genes as it seemed structurally and functionally more complex than other simpler organisms. This along with the belief of “one gene, one protein”, were demonstrated to be incorrect. The inequality in the ratio of gene to protein formation gave rise to the theory of alternative splicing (AS). AS is a mechanism by which one gene gives rise to multiple protein products. Numerous databases and online bioinformatic tools are available for the detection and analysis of AS. Bioinformatics provides an important approach to study mRNA and protein diversity by various tools such as expressed sequence tag (EST) sequences obtained from completely processed mRNA. Microarrays and deep sequencing approaches also aid in the detection of splicing events. Initially it was postulated that AS occurred only in about 5%; of all genes but was later found to be more abundant. Using bioinformatic approaches, the level of AS in human genes was found to be fairly high with 35-59%; of genes having at least one AS form. Our ability to determine and predict AS is important as disorders in splicing patterns may lead to abnormal splice variants resulting in genetic diseases. In addition, the diversity of proteins produced by AS poses a challenge for successful drug discovery and therefore a greater understanding of AS would be beneficial.

Analysis of organic aerosols collected on filters by Aerosol Mass Spectrometry for source identification

Aerosol mass spectrometers (AMS) are powerful tools in the analysis of the chemical composition of airborne particles, particularly organic aerosols which are gaining increasing attention. However, the advantages of AMS in providing on-line data can be outweighed by the difficulties involved in its use in field measurements at multiple sites. In contrast to the on-line measurement by AMS, a method which involves sample collection on filters followed by subsequent analysis by AMS could significantly broaden the scope of AMS application. We report the application of such an approach to field studies at multiple sites. An AMS was deployed at 5 urban schools to determine the sources of the organic aerosols at the schools directly. PM1 aerosols were also collected on filters at these and 20 other urban schools. The filters were extracted with water and the extract run through a nebulizer to generate the aerosols, which were analysed by an AMS. The mass spectra from the samples collected on filters at the 5 schools were found to have excellent correlations with those obtained directly by AMS, with r2 ranging from 0.89 to 0.98. Filter recoveries varied between the schools from 40 -115%, possibly indicating that this method provides qualitative rather than quantitative information. The stability of the organic aerosols on Teflon filters was demonstrated by analysing samples stored for up to two years. Application of the procedure to the remaining 20 schools showed that secondary organic aerosols were the main source of aerosols at the majority of the schools. Overall, this procedure provides accurate representation of the mass spectra of ambient organic aerosols and could facilitate rapid data acquisition at multiple sites where AMS could not be deployed for logistical reasons.

A single dose of azithromycin does not improve clinical outcomes of children hospitalised with bronchiolitis : a randomised, placebo-controlled trial

Objective Bronchiolitis, one of the most common reasons for hospitalisation in young children, is particularly problematic in Indigenous children. Macrolides may be beneficial in settings where children have high rates of nasopharyngeal bacterial carriage and frequent prolonged illness. The aim of our double-blind placebo-controlled randomised trial was to determine if a large single dose of azithromycin (compared to placebo) reduced length of stay (LOS), duration of oxygen (O2) and respiratory readmissions within 6 months of children hospitalised with bronchiolitis. We also determined the effect of azithromycin on nasopharyngeal microbiology. Methods Children aged ≤18 months were randomised to receive a single large dose (30 mg/kg) of either azithromycin or placebo within 24 hrs of hospitalisation. Nasopharyngeal swabs were collected at baseline and 48hrs later. Primary endpoints (LOS, O2) were monitored every 12 hrs. Hospitalised respiratory readmissions 6-months post discharge was collected. Results 97 children were randomised (n = 50 azithromycin, n = 47 placebo). Median LOS was similar in both groups; azithromycin = 54 hours, placebo = 58 hours (difference between groups of 4 hours 95%CI -8, 13, p = 0.6). O2 requirement was not significantly different between groups; Azithromycin = 35 hrs; placebo = 42 hrs (difference 7 hours, 95%CI -9, 13, p = 0.7). Number of children re-hospitalised was similar 10 per group (OR = 0.9, 95%CI 0.3, 2, p = 0.8). At least one virus was detected in 74% of children. The azithromycin group had reduced nasopharyngeal bacterial carriage (p = 0.01) but no difference in viral detection at 48 hours. Conclusion Although a single dose of azithromycin reduces carriage of bacteria, it is unlikely to be beneficial in reducing LOS, duration of O2 requirement or readmissions in children hospitalised with bronchiolitis. It remains uncertain if an earlier and/or longer duration of azithromycin improves clinical and microbiological outcomes for children.

Design, construction, and analysis of cell line arrays and tissue microarrays for gene expression analysis

Cell line array (CMA) and tissue microarray (TMA) technologies are high-throughput methods for analysing both the abundance and distribution of gene expression in a panel of cell lines or multiple tissue specimens in an efficient and cost-effective manner. The process is based on Kononen's method of extracting a cylindrical core of paraffin-embedded donor tissue and inserting it into a recipient paraffin block. Donor tissue from surgically resected paraffin-embedded tissue blocks, frozen needle biopsies or cell line pellets can all be arrayed in the recipient block. The representative area of interest is identified and circled on a haematoxylin and eosin (H&E)-stained section of the donor block. Using a predesigned map showing a precise spacing pattern, a high density array of up to 1,000 cores of cell pellets and/or donor tissue can be embedded into the recipient block using a tissue arrayer from Beecher Instruments. Depending on the depth of the cell line/tissue removed from the donor block 100-300 consecutive sections can be cut from each CMA/TMA block. Sections can be stained for in situ detection of protein, DNA or RNA targets using immunohistochemistry (IHC), fluorescent in situ hybridisation (FISH) or mRNA in situ hybridisation (RNA-ISH), respectively. This chapter provides detailed methods for CMA/TMA design, construction and analysis with in-depth notes on all technical aspects including tips to deal with common pitfalls the user may encounter. © Springer Science+Business Media, LLC 2011.

Consequences of introducing mandatory reporting legislation for child sexual abuse

To improve detection of child sexual abuse, many jurisdictions have enacted mandatory reporting laws requiring selected persons to report known and suspected cases. In Ireland, the Child First approach previously incorporated only a policy-based approach to reporting. Due to a perceived lack of efficacy, the Children First Bill was drafted in 2012 to shift this policy guidance to a legislative approach. What effects will the new legislative reporting duties have on numbers of reports, and outcomes of reports, of suspected child sexual abuse? This paper will shed light on these important questions by presenting results of analyses of the introduction of legislative reporting obligations in two Australian States. Three questions will be explored: 1. Does introducing reporting legislation result in enhanced detection of child sexual abuse? 2. Do different reporter groups have different patterns of reporting? 3. What do the patterns of report numbers and outcomes indicate for child protection systems and communities?

Error modeling and calibration of exteroceptive sensors for accurate mapping applications

Reliable robotic perception and planning are critical to performing autonomous actions in uncertain, unstructured environments. In field robotic systems, automation is achieved by interpreting exteroceptive sensor information to infer something about the world. This is then mapped to provide a consistent spatial context, so that actions can be planned around the predicted future interaction of the robot and the world. The whole system is as reliable as the weakest link in this chain. In this paper, the term mapping is used broadly to describe the transformation of range-based exteroceptive sensor data (such as LIDAR or stereo vision) to a fixed navigation frame, so that it can be used to form an internal representation of the environment. The coordinate transformation from the sensor frame to the navigation frame is analyzed to produce a spatial error model that captures the dominant geometric and temporal sources of mapping error. This allows the mapping accuracy to be calculated at run time. A generic extrinsic calibration method for exteroceptive range-based sensors is then presented to determine the sensor location and orientation. This allows systematic errors in individual sensors to be minimized, and when multiple sensors are used, it minimizes the systematic contradiction between them to enable reliable multisensor data fusion. The mathematical derivations at the core of this model are not particularly novel or complicated, but the rigorous analysis and application to field robotics seems to be largely absent from the literature to date. The techniques in this paper are simple to implement, and they offer a significant improvement to the accuracy, precision, and integrity of mapped information. Consequently, they should be employed whenever maps are formed from range-based exteroceptive sensor data. © 2009 Wiley Periodicals, Inc.

Typology of phishing email victims based on their behavioural response

A victim of phishing emails could be subjected to money loss and identity theft. This paper investigates the different types of phishing email victims, with the goal of increasing such victims' defences. To obtain this kind of information, an experiment which involves sending a phishing email to participants is conducted. Quantitative and qualitative methods are also used to collect users' information. A model for detecting deception has been employed to understand victims' behaviour. This paper reports the qualitative results. The findings suggest that victims of phishing emails do not always exhibit the same vulnerability. The cause of being a victim is a result of three weaknesses in the detection process: (1) lack of knowledge; (2) weak confirmation channel, and; (3) victims' high propensity towards risk-taking. Therefore, it is suggested that users be provided with suitable confirmation channels and be more risk averse in their behaviour so that they would not fall victim to phishing emails.