Rescaling clustering trees using impact ratios for robust hierarchical speaker clustering

We present a novel method for improving hierarchical speaker clustering in the tasks of speaker diarization and speaker linking. In hierarchical clustering, a tree can be formed that demonstrates various levels of clustering. We propose a ratio that expresses the impact of each cluster on the formation of this tree and use this to rescale cluster scores. This provides score normalisation based on the impact of each cluster. We use a state-of-the-art speaker diarization and linking system across the SAIVT-BNEWS corpus to show that our proposed impact ratio can provide a relative improvement of 16% in diarization error rate (DER).

Physical activity, walking and quality of life in women with depressive symptoms

Background Physical activity (PA) has a positive association with health-related quality of life (HRQL) in the general population. The association between PA and HRQL in those with poor mental health is less clear. Purpose To examine the concurrent and prospective dose-response relationships between total physical activity (TPA) and walking only with HRQL in women aged 50-55 with depressive symptoms in 2001. Methods Participants were 1904 women born in 1946-1951 who completed mailed surveys for the Australian Longitudinal Study on Women's Health in 2001, 2004, 2007 and 2010 and who, in 2001, reported depressive symptoms. At each time point, they reported their weekly minutes of walking, moderate PA, and vigorous PA. A summary TPA score was created that accounted for differences in energy expenditure among the three PA types. Mixed models were used to examine associations between TPA and HRQL (SF-36 component and subscale scores) and between walking and HRQL, for women who reported walking as their only PA. Analyses were conducted in 2013-2014. Results Concurrently, higher levels of TPA and walking were associated with better HRQL (p<0.05). The strongest associations were found for physical functioning, vitality, and social functioning subscales. In prospective models, associations were attenuated, yet compared with women doing no TPA or walking, women doing “sufficient” TPA or walking had significantly better HRQL over time for most SF-36 scales. Conclusions This study extends previous work by demonstrating trends between both TPA and walking and HRQL in women reporting depressive symptoms.

A normal-distribution based reputation model

Rating systems are used by many websites, which allow customers to rate available items according to their own experience. Subsequently, reputation models are used to aggregate available ratings in order to generate reputation scores for items. A problem with current reputation models is that they provide solutions to enhance accuracy of sparse datasets not thinking of their models performance over dense datasets. In this paper, we propose a novel reputation model to generate more accurate reputation scores for items using any dataset; whether it is dense or sparse. Our proposed model is described as a weighted average method, where the weights are generated using the normal distribution. Experiments show promising results for the proposed model over state-of-the-art ones on sparse and dense datasets.

A rating aggregation method for generating product reputations

Many websites offer the opportunity for customers to rate items and then use customers' ratings to generate items reputation, which can be used later by other users for decision making purposes. The aggregated value of the ratings per item represents the reputation of this item. The accuracy of the reputation scores is important as it is used to rank items. Most of the aggregation methods didn't consider the frequency of distinct ratings and they didn't test how accurate their reputation scores over different datasets with different sparsity. In this work we propose a new aggregation method which can be described as a weighted average, where weights are generated using the normal distribution. The evaluation result shows that the proposed method outperforms state-of-the-art methods over different sparsity datasets.

After the flood

FRUSTRATED residents of Grantham, the Lockyer Valley township devastated by Queensland's deadly summer floods, are demanding that the commission of inquiry into the disaster investigate whether an earth wall around a sand quarry helped cause the "inland tsunami" that killed 12 people and destroyed scores of homes.

Using the Driver Behaviour Questionnaire to predict crashes and demerit point loss : does it get better with larger sample sizes?

The Driver Behaviour Questionnaire (DBQ) continues to be the most widely utilised self-report scale globally to assess crash risk and aberrant driving behaviours among motorists. However, the scale also attracts criticism regarding its perceived limited ability to accurately identify those most at risk of crash involvement. This study reports on the utilisation of the DBQ to examine the self-reported driving behaviours (and crash outcomes) of drivers in three separate Australian fleet samples (N = 443, N = 3414, & N = 4792), and whether combining the samples increases the tool’s predictive ability. Either on-line or paper versions of the questionnaire were completed by fleet employees in three organisations. Factor analytic techniques identified either three or four factor solutions (in each of the separate studies) and the combined sample produced expected factors of: (a) errors, (b) highway-code violations and (c) aggressive driving violations. Highway code violations (and mean scores) were comparable across the studies. However, across the three samples, multivariate analyses revealed that exposure to the road was the best predictor of crash involvement at work, rather than DBQ constructs. Furthermore, combining the scores to produce a sample of 8649 drivers did not improve the predictive ability of the tool for identifying crashes (e.g., 0.4% correctly identified) or for demerit point loss (0.3%). The paper outlines the major findings of this comparative sample study in regards to utilising self-report measurement tools to identify “at risk” drivers as well as the application of such data to future research endeavours.

Evidence of associations between cytokine gene polymorphisms and quality of life in patients with cancer and their family caregivers

PURPOSE/OBJECTIVES: To identify latent classes of individuals with distinct quality-of-life (QOL) trajectories, to evaluate for differences in demographic characteristics between the latent classes, and to evaluate for variations in pro- and anti-inflammatory cytokine genes between the latent classes. DESIGN: Descriptive, longitudinal study. SETTING: Two radiation therapy departments located in a comprehensive cancer center and a community-based oncology program in northern California. SAMPLE: 168 outpatients with prostate, breast, brain, or lung cancer and 85 of their family caregivers (FCs). METHODS: Growth mixture modeling (GMM) was employed to identify latent classes of individuals based on QOL scores measured prior to, during, and for four months following completion of radiation therapy. Single nucleotide polymorphisms (SNPs) and haplotypes in 16 candidate cytokine genes were tested between the latent classes. Logistic regression was used to evaluate the relationships among genotypic and phenotypic characteristics and QOL GMM group membership. MAIN RESEARCH VARIABLES: QOL latent class membership and variations in cytokine genes. FINDINGS: Two latent QOL classes were found: higher and lower. Patients and FCs who were younger, identified with an ethnic minority group, had poorer functional status, or had children living at home were more likely to belong to the lower QOL class. After controlling for significant covariates, between-group differences were found in SNPs in interleukin 1 receptor 2 (IL1R2) and nuclear factor kappa beta 2 (NFKB2). For IL1R2, carrying one or two doses of the rare C allele was associated with decreased odds of belonging to the lower QOL class. For NFKB2, carriers with two doses of the rare G allele were more likely to belong to the lower QOL class. CONCLUSIONS: Unique genetic markers in cytokine genes may partially explain interindividual variability in QOL. IMPLICATIONS FOR NURSING: Determination of high-risk characteristics and unique genetic markers would allow for earlier identification of patients with cancer and FCs at higher risk for poorer QOL. Knowledge of these risk factors could assist in the development of more targeted clinical or supportive care interventions for those identified.

An evidence-based program to improve analgesic practice and pain outcomes in residential aged care facilities

Pain is common in individuals living in residential aged care facilities (RACFs), and a number of obstacles have been identified as recurring barriers to adequate pain management. To address this, the Australian Pain Society developed 27 recommendations for comprehensive good practice in the identification, assessment, and management of pain. This study reviewed preexisting pain management practice at five Australian RACFs and identified changes needed to implement the recommendations and then implemented an evidence-based program that aimed to facilitate better pain management. The program involved staff training and education and revised in-house pain-management procedures. Reviews occurred before and after the program and included the assessment of 282 residents for analgesic use and pain status. Analgesic use improved after the program (P<.001), with a decrease in residents receiving no analgesics (from 15% to 6%) and an increase in residents receiving around-the-clock plus as-needed analgesics (from 24% to 43%). There were improvements in pain relief for residents with scores indicative of pain, with Abbey pain scale (P=.005), Pain Assessment in Advanced Dementia Scale (P=.001), and Non-communicative Patient's Pain Assessment Instrument scale (P<.001) scores all improving. Although physical function declined as expected, Medical Outcomes Study 36-item Short-Form Survey bodily pain scores also showed improvement (P=.001). Better evidence-based practice and outcomes in RACFs can be achieved with appropriate training and education. Investing resources in the aged care workforce using this program improved analgesic practice and pain relief in participating sites. Further attention to the continued targeted pain management training of aged care staff is likely to improve pain-focused care for residents.

Observing the mother-infant feeding interaction

Early parenting is critical to effective attachment and a range of positive developmental outcomes for children. Feeding is a key task of early parenting and increasing evidence indicates that early feeding practices are important for the development of self-regulation of intake and food preferences which in turn are predictors of later obesity risk. However, relatively little is known about the mother-infant interaction at the transition to solids among typically developing children. This study aimed to describe parenting strategies used by mothers at the transition from milk feeding to solid food. Twenty mother-infant dyads were video-taped during a feeding interaction and data was analysed to describe maternal use of parenting strategies. It was predicted that positive feeding strategies would be correlated with lower levels of Infant Food Refusal (IFR), higher maternal sensitivity, and better overall parenting scores. The opposite was predicted for negative feeding strategies. It was found that positive strategy use and general parenting scores were significantly correlated in the predicted direction, however maternal instruction, aversive contact and ineffective strategy use were significantly correlated with and predictive of IFR. Additionally, it was hypothesised that maternal strategy use would deteriorate towards the end of the interaction, and this hypothesis was partially supported: significantly more negative strategy use was observed in the last third of the interaction, whilst positive strategy use remained consistent through the feeding interaction. The findings have important implications for early feeding parent education and intervention programs.

Can mothers accurately describe their toddler’s weight? Results from the NOURISH cohort

A mother’s perception of her child’s weight may be more important in determining how she feeds her child, than the child’s actual weight status. Use of controlling feeding practices, prompted by perceptions and concerns about weight, may disrupt the child’s innate self-regulation of energy intake. This can promote overeating and overweight (Costanzo & Woody, 1985). This study describes mother’s perception of her child’s weight relative to the child’s actual weight. Mothers in the control group of NOURISH (n=276) were asked to describe their child as underweight, normal weight, or somewhat/very overweight via self-administered questionnaire when children were aged 12-16 months (Daniels et al, 2009). Child’s weight and length were measured by study staff. At assessment, mean age (sd) was 13.7(1.3) months, mean weight-for-age z-score (sd) was 0.6(0.8) (WHO standards, 2008), and 51% were male. Twenty-seven children were perceived as underweight (10%) and twelve children were perceived as overweight (4%). ANOVA revealed significant differences in weight-for-age z-scores across each category of weight perception, mean (sd) -0.2(0.5), 0.6(0.8) and 1.8(0.7) for underweight, normal weight and overweight respectively F(4, 288)= 15.6, (p<0.00). Based on WHO criteria only one of the 27 children was correctly perceived as underweight (WHO 2008). Similarly while 12 children were perceived as overweight, 88 were actually overweight/at risk. At group level, children of mothers who perceived their child as underweight were indeed leaner. However at the individual level mothers could not accurately describe their child’s weight, tending to over-identify underweight and perceive overweight children as normal weight.

A tag-based personalized item recommendation system using tensor modeling and topic model approaches

This research falls in the area of enhancing the quality of tag-based item recommendation systems. It aims to achieve this by employing a multi-dimensional user profile approach and by analyzing the semantic aspects of tags. Tag-based recommender systems have two characteristics that need to be carefully studied in order to build a reliable system. Firstly, the multi-dimensional correlation, called as tag assignment , should be appropriately modelled in order to create the user profiles [1]. Secondly, the semantics behind the tags should be considered properly as the flexibility with their design can cause semantic problems such as synonymy and polysemy [2]. This research proposes to address these two challenges for building a tag-based item recommendation system by employing tensor modeling as the multi-dimensional user profile approach, and the topic model as the semantic analysis approach. The first objective is to optimize the tensor model reconstruction and to improve the model performance in generating quality rec-ommendation. A novel Tensor-based Recommendation using Probabilistic Ranking (TRPR) method [3] has been developed. Results show this method to be scalable for large datasets and outperforming the benchmarking methods in terms of accuracy. The memory efficient loop implements the n-mode block-striped (matrix) product for tensor reconstruction as an approximation of the initial tensor. The probabilistic ranking calculates the probabil-ity of users to select candidate items using their tag preference list based on the entries generated from the reconstructed tensor. The second objective is to analyse the tag semantics and utilize the outcome in building the tensor model. This research proposes to investigate the problem using topic model approach to keep the tags nature as the “social vocabulary” [4]. For the tag assignment data, topics can be generated from the occurrences of tags given for an item. However there is only limited amount of tags availa-ble to represent items as collection of topics, since an item might have only been tagged by using several tags. Consequently, the generated topics might not able to represent the items appropriately. Furthermore, given that each tag can belong to any topics with various probability scores, the occurrence of tags cannot simply be mapped by the topics to build the tensor model. A standard weighting technique will not appropriately calculate the value of tagging activity since it will define the context of an item using a tag instead of a topic.

Testing the effectiveness of a self-efficacy based exercise intervention for adults with venous leg ulcers : protocol of a randomised controlled trial

Background: Exercise and adequate self-management capacity may be important strategies in the management of venous leg ulcers. However, it remains unclear if exercise improves the healing rates of venous leg ulcers and if a self-management exercise program based on self-efficacy theory is well adhered to. Method/Design: This is a randomised controlled in adults with venous leg ulcers to determine the effectiveness of a self-efficacy based exercise intervention. Participants with venous leg ulcers are recruited from 3 clinical sites in Australia. After collection of baseline data, participants are randomised to either an intervention group or control group. The control group receive usual care, as recommended by evidence based guidelines. The intervention group receive an individualised program of calf muscle exercises and walking. The twelve week exercise program integrates multiple elements, including up to six telephone delivered behavioural coaching and goal setting sessions, supported by written materials, a pedometer and two follow-up booster calls if required. Participants are encouraged to seek social support among their friends, self-monitor their weekly steps and lower limb exercises. The control group are supported by a generic information sheet that the intervention group also receive encouraging lower limb exercises, a pedometer for self-management and phone calls at the same time points as the intervention group. The primary outcome is the healing rates of venous leg ulcers which are assessed at fortnightly clinic appointments. Secondary outcomes, assessed at baseline and 12 weeks: functional ability (range of ankle motion and Tinetti gait and balance score), quality of life and self-management scores. Discussion: This study seeks to address a significant gap in current wound management practice by providing evidence for the effectiveness of a home-based exercise program for adults with venous leg ulcers. Theory-driven, evidence-based strategies that can improve an individual’s exercise self-efficacy and self-management capacity could have a significant impact in improving the management of people with venous leg ulcers. Information gained from this study will provide much needed information on management of this chronic disease to promote health and independence in this population. Trial registration: Australian New Zealand Clinical Trials Registry ACTRN12612000475842 Trial status: Current follow up

An MRF based abnormal event detection approach using motion and appearance features

Abnormal event detection has attracted a lot of attention in the computer vision research community during recent years due to the increased focus on automated surveillance systems to improve security in public places. Due to the scarcity of training data and the definition of an abnormality being dependent on context, abnormal event detection is generally formulated as a data-driven approach where activities are modeled in an unsupervised fashion during the training phase. In this work, we use a Gaussian mixture model (GMM) to cluster the activities during the training phase, and propose a Gaussian mixture model based Markov random field (GMM-MRF) to estimate the likelihood scores of new videos in the testing phase. Further-more, we propose two new features: optical acceleration, and the histogram of optical flow gradients; to detect the presence of any abnormal objects and speed violations in the scene. We show that our proposed method outperforms other state of the art abnormal event detection algorithms on publicly available UCSD dataset.

Integrating ‘role play’ in assessment to strengthen professional conduct and accountability of students : a pilot study

BACKGROUND In a process engineering setting, graduates are frequently allocated reviews of existing operations or required to scope new production processes by their supervisors with a view to improving or expanding on operations and overall productivity. These tasks may be carried out in teams and in consultation with the process engineer’s immediate line manager or a more experienced engineer, such as the Production or Maintenance Manager; ultimately reporting to senior management, which is frequently a non-engineer. Although professional skills development is part of engineering curricula, ‘professional conduct’ and ‘accountability’ required for dealing with peers and superiors in industry is not very well addressed at university. Consequently, upon graduation, many students are, in terms of knowledge and experience in this area, underprepared to work effectively in industry settings. PURPOSE The purpose of this study was to develop and implement a role-play scenario within a core 2nd year process engineering unit, so that students could gain knowledge, skills and experience in different aspects (and nuances) of professional conduct and accountability. DESIGN/METHOD In the role-play scenario, students worked in ‘engineering production teams’ to design a process for an iconic Queensland fruitcake and to present their solution and recommendations (culminating in a poster presentation) to an assessment panel consisting of staff, role-playing as, ‘production and plant managers’. Students were assessed on several areas, including professionalism using a criteria referenced assessment guide by a 3-member cross-disciplinary staff panel consisting of a Business Faculty lecturer, an engineer from industry and the lecturer of the Process Engineering unit. Professional conduct and accountability was gauged through direct questioning by the panel. Feedback was also sought from students on various aspects through a survey questionnaire after the role play activity at the end of semester. RESULTS Overall, the role play was very well performed with students achieving an average score of 79.3/100 (distinction grade). Professional conduct as assessed by panel was on average better than scores given for professional accountability (4.0 compared with 3.6 out of 5). Feedback from students indicated that the learning activities had contributed to their overall understanding of the content and the role of process engineers. Industry involvement was rated very highly as contributing to their learning at 4.8 (on Likert scale from 1 – 5) and the poster presentation was rated at 3.6. CONCLUSIONS This pilot study was successful in implementing a new assessment task for modelling professional conduct and accountability within a 2nd year core unit. This task incorporated a role-play activity and there was evidence to suggest that this and associated learning tasks were successful in broadening students’ understanding and skills in this area required for engineering practice. Following feedback given by students and staff, improvements will be made to the nature of the problem, how it is defined, its assessment, and the approach taken in the role-play scenario when the unit is offered in 2014.

The effect of work-integrated learning in highlighting the complexity of work : a pilot study of work self-efficacy

This pilot study aims to examine the effect of work-integrated learning (WIL) on work self-efficacy (WSE) for undergraduate students from the Queensland University of Technology. A WSE instrument was used to examine the seven subscales of WSE. These were; learning, problem solving, pressure, role expectations, team work, sensitivity and work politics. The results of this pilot study revealed that, overall the WSE scores were highest when the students’ did not participate in the WIL unit (comparison group) in comparison to the WIL group. The current paper suggests that WSE scores were changed as a result of WIL participation. These findings open a new path for future studies allowing them to explore the relationship between WIL and the specific subscales of WSE.