Topic model kernel : an empirical study towards probabilistically reduced features for classification

Probabilistic topic models have become a standard in modern machine learning with wide applications in organizing and summarizing ‘documents’ in high-dimensional data such as images, videos, texts, gene expression data, and so on. Representing data by dimensional reduction of mixture proportion extracted from topic models is not only richer in semantics than bag-of-word interpretation, but also more informative for classification tasks. This paper describes the Topic Model Kernel (TMK), a high dimensional mapping for Support Vector Machine classification of data generated from probabilistic topic models. The applicability of our proposed kernel is demonstrated in several classification tasks from real world datasets. We outperform existing kernels on the distributional features and give the comparative results on non-probabilistic data types.

Investigating machine learning techniques for detection of depression using structural MRI volumetric features

Structural MRI offers anatomical details and high sensitivity to pathological changes. It can demonstrate certain patterns of brain changes present at a structural level. Research to date has shown that volumetric analysis of brain regions has importance in depression detection. However, such analysis has had very minimal use in depression detection studies at individual level. Optimally combining various brain volumetric features/attributes, and summarizing the data into a distinctive set of variables remain difficult. This study investigates machine learning algorithms that automatically identify relevant data attributes for depression detection. Different machine learning techniques are studied for depression classification based on attributes extracted from structural MRI (sMRI) data. The attributes include volume calculated from whole brain, white matter, grey matter and hippocampus. Attributes subset selection is performed aiming to remove redundant attributes using three filtering methods and one hybrid method, in combination with ranker search algorithms. The highest average classification accuracy, obtained by using a combination of both SVM-EM and IG-Random Tree algorithms, is 85.23%. The classification approach implemented in this study can achieve higher accuracy than most reported studies using sMRI data, specifically for detection of depression.

Clinical features and risk factors of oedematous Mycobacterium ulcerans lesions in an Australian population: beware cellulitis in an endemic area

Introduction
Oedematous lesions are a less common but more severe form of Mycobacterium ulcerans disease. Misdiagnosis as bacterial cellulitis can lead to delays in treatment. We report the first comprehensive descriptions of the clinical features and risk factors of patients with oedematous disease from the Bellarine Peninsula of south-eastern Victoria, Australia.

Methods
Data on all confirmed Mycobacterium ulcerans cases managed at Barwon Health, Victoria, were collected from 1/1/1998–31/12/2012. A multivariate logistic regression model was used to assess associations with oedematous forms of Mycobacterium ulcerans disease.

Results
Seventeen of 238 (7%) patients had oedematous Mycobacterium ulcerans lesions. Their median age was 70 years (IQR 17–82 years) and 71% were male. Twenty-one percent of lesions were WHO category one, 35% category two and 41% category three. 16 (94%) patients were initially diagnosed with cellulitis and received a median 14 days (IQR 9–17 days) of antibiotics and 65% required hospitalization prior to Mycobacterium ulcerans diagnosis. Fever was present in 50% and pain in 87% of patients. The WCC, neutrophil count and CRP were elevated in 54%, 62% and 75% of cases respectively. The median duration of antibiotic treatment was 84 days (IQR 67–96) and 94% of cases required surgical intervention. On multivariable analysis, there was an increased likelihood of a lesion being oedematous if on the hand (OR 85.62, 95% CI 13.69–535.70; P<0.001), elbow (OR 7.83, 95% CI 1.39–43.96; p<0.001) or ankle (OR 7.92, 95% CI 1.28–49.16; p<0.001), or if the patient had diabetes mellitus (OR 9.42, 95% CI 1.62–54.74; p = 0.02).

Conclusions
In an Australian population, oedematous Mycobacterium ulcerans lesions present with similar symptoms, signs and investigation results to, and are commonly mistakenly diagnosed for, bacterial limb cellulitis. There is an increased likelihood of oedematous lesions affecting the hand, elbow or ankle, and in patients with diabetes.

Complexities of teaching — features of teacher education programs and professional learning practices to support their development

This publication is the editorial for Vol 15.2 of MTED which summarises and synthesises the articles published in the issue. The extent to which those whose learning is the focus controlled the learning situations in which they participated is the key concept discussed and the factors contributing to the extent of control are elucidated.

A cross-cultural study investigating body features associated with male adolescents' body dissatisfaction in Australia, China, and Malaysia

This study investigated how dissatisfaction with particular aspects of the body was associated with overall body dissatisfaction among male adolescents in Western and Asian cultures. One hundred and six Malaysian Malays, 55 Malaysian Chinese, 195 Chinese from China, and 45 non-Asian Australians aged 12 to 19 years completed a questionnaire assessing dissatisfaction with their overall body and dissatisfaction with varying aspects of their body. Dissatisfaction with the face, height, and hair was positively correlated with overall body dissatisfaction among Malaysian Malays after body mass index, age and dissatisfaction with body areas typically included in measures (weight/shape, upper, middle, and lower body, and muscles) had been controlled for. Dissatisfaction with the face was positively correlated with overall body dissatisfaction among Malaysian Chinese. These findings demonstrate the differences in body focus for males from different cultures and the importance of using assessment measures that address all possible areas of body focus.

Using manipulated photographs to identify features of streetscapes that may encourage older adults to walk for transport

Experimental evidence of environmental features important for physical activity is challenging to procure in real world settings. The current study aimed to investigate the causal effects of environmental modifications on a photographed street's appeal for older adults' walking for transport. Secondly, we examined whether these effects differed according to gender, functional limitations, and current level of walking for transport. Thirdly, we examined whether different environmental modifications interacted with each other. Qualitative responses were also reported to gain deeper insight into the observed quantitative relationships. Two sets of 16 panoramic photographs of a streetscape were created, in which six environmental factors were manipulated (sidewalk evenness, traffic level, general upkeep, vegetation, separation from traffic, and benches). Sixty older adults sorted these photographs on appeal for walking for transport on a 7-point scale and reported qualitative information on the reasons for their rankings. Sidewalk evenness appeared to have the strongest influence on a street's appeal for transport-related walking. The effect of sidewalk evenness was even stronger when the street's overall upkeep was good and when traffic was absent. Absence of traffic, presence of vegetation, and separation from traffic also increased a street's appeal for walking for transport. There were no moderating effects by gender or functional limitations. The presence of benches increased the streetscape's appeal among participants who already walked for transport at least an hour/week. The protocols and methods used in the current study carry the potential to further our understanding of environment-PA relationships. Our findings indicated sidewalk evenness as the most important environmental factor influencing a street's appeal for walking for transport among older adults. However, future research in larger samples and in real-life settings is needed to confirm current findings.

Sex differences in the Toll-like receptor-mediated response of plasmacytoid dendritic cells to HIV-1

Manifestations of viral infections can differ between women and men, and marked sex differences have been described in the course of HIV-1 disease. HIV-1-infected women tend to have lower viral loads early in HIV-1 infection but progress faster to AIDS for a given viral load than men. Here we show substantial sex differences in the response of plasmacytoid dendritic cells (pDCs) to HIV-1. pDCs derived from women produce markedly more interferon-alpha (IFN-alpha) in response to HIV-1-encoded Toll-like receptor 7 (TLR7) ligands than pDCs derived from men, resulting in stronger secondary activation of CD8(+) T cells. In line with these in vitro studies, treatment-naive women chronically infected with HIV-1 had considerably higher levels of CD8(+) T cell activation than men after adjusting for viral load. These data show that sex differences in TLR-mediated activation of pDCs may account for higher immune activation in women compared to men at a given HIV-1 viral load and provide a mechanism by which the same level of viral replication might result in faster HIV-1 disease progression in women compared to men. Modulation of the TLR7 pathway in pDCs may therefore represent a new approach to reduce HIV-1-associated pathology.

Degree of contribution (DoC) feature selection algorithm for structural brain MRI volumetric features in depression detection

Accurate detection of depression at an individual level using structural magnetic resonance imaging (sMRI) remains a challenge. Brain volumetric changes at a structural level appear to have importance in depression biomarkers studies. An automated algorithm is developed to select brain sMRI volumetric features for the detection of depression.