A matrix factorization framework for jointly analyzing multiple nonnegative data source

Nonnegative matrix factorization based methods provide one of the simplest and most effective approaches to text mining. However, their applicability is mainly limited to analyzing a single data source. In this paper, we propose a novel joint matrix factorization framework which can jointly analyze multiple data sources by exploiting their shared and individual structures. The proposed framework is flexible to handle any arbitrary sharing configurations encountered in real world data. We derive an efficient algorithm for learning the factorization and show that its convergence is theoretically guaranteed. We demonstrate the utility and effectiveness of the proposed framework in two real-world applications–improving social media retrieval using auxiliary sources and cross-social media retrieval. Representing each social media source using their textual tags, for both applications, we show that retrieval performance exceeds the existing state-of-the-art techniques. The proposed solution provides a generic framework and can be applicable to a wider context in data mining wherever one needs to exploit mutual and individual knowledge present across multiple data sources.

A Bayesian framework for learning shared and individual subspaces from multiple data sources

This paper presents a novel Bayesian formulation to exploit shared structures across multiple data sources, constructing foundations for effective mining and retrieval across disparate domains. We jointly analyze diverse data sources using a unifying piece of metadata (textual tags). We propose a method based on Bayesian Probabilistic Matrix Factorization (BPMF) which is able to explicitly model the partial knowledge common to the datasets using shared subspaces and the knowledge specific to each dataset using individual subspaces. For the proposed model, we derive an efficient algorithm for learning the joint factorization based on Gibbs sampling. The effectiveness of the model is demonstrated by social media retrieval tasks across single and multiple media. The proposed solution is applicable to a wider context, providing a formal framework suitable for exploiting individual as well as mutual knowledge present across heterogeneous data sources of many kinds.

Modelling multilevel data in multimedia : A hierarchical factor analysis approach

Multimedia content understanding research requires rigorous approach to deal with the complexity of the data. At the crux of this problem is the method to deal with multilevel data whose structure exists at multiple scales and across data sources. A common example is modeling tags jointly with images to improve retrieval, classification and tag recommendation. Associated contextual observation, such as metadata, is rich that can be exploited for content analysis. A major challenge is the need for a principal approach to systematically incorporate associated media with the primary data source of interest. Taking a factor modeling approach, we propose a framework that can discover low-dimensional structures for a primary data source together with other associated information. We cast this task as a subspace learning problem under the framework of Bayesian nonparametrics and thus the subspace dimensionality and the number of clusters are automatically learnt from data instead of setting these parameters a priori. Using Beta processes as the building block, we construct random measures in a hierarchical structure to generate multiple data sources and capture their shared statistical at the same time. The model parameters are inferred efficiently using a novel combination of Gibbs and slice sampling. We demonstrate the applicability of the proposed model in three applications: image retrieval, automatic tag recommendation and image classification. Experiments using two real-world datasets show that our approach outperforms various state-of-the-art related methods.

A matrix factorization framework for jointly analyzing multiple nonnegative data sources

Nonnegative matrix factorization based methods provide one of the simplest and most effective approaches to text mining. However, their applicability is mainly limited to analyzing a single data source. In this chapter, we propose a novel joint matrix factorization framework which can jointly analyze multiple data sources by exploiting their shared and individual structures. The proposed framework is flexible to handle any arbitrary sharing configurations encountered in real world data. We derive an efficient algorithm for learning the factorization and show that its convergence is theoretically guaranteed. We demonstrate the utility and effectiveness of the proposed framework in two real-world applications—improving social media retrieval using auxiliary sources and cross-social media retrieval. Representing each social media source using their textual tags, for both applications, we show that retrieval performance exceeds the existing state-of-the-art techniques. The proposed solution provides a generic framework and can be applicable to a wider context in data mining wherever one needs to exploit mutual and individual knowledge present across multiple data sources.

Corporate financial disclosure practices in Asian and Middle Eastern countries

This paper investigates the financial disclosure practices of corporate annual reports published in Asian countries including Bangladesh, Indonesian, Malaysia and the Middle East countries including Bahrain, Iran, Jordan, Kuwait, Oman, Pakistan, Qatar, Saudi Arabia and Turkey. The purpose of the study is to measure the financial disclosure diversity in these countries, with a view to developing a classification of their similarities and differences in respect to their compliance with International Accounting Standards (IAS). Annual reports of 126 public companies liisted on the countries' stock exchanges are the central data source, supplemented with other relevant information about financial disclosure practices in each country. A disclosure checklist adopted from all IASs and summarised in 306 individual items of financial disclosures is used as a means of extending an understanding of financial reporting in these countries. Results show the relative degree of conformity with IASs for each of the countries included in this study. 

The hazards of starting and quitting smoking: some Australian evidence

The empirical analysis employs individual level data from the Australian Health Survey combined with retrospective data on tobacco price matched to the age at which the individual started and quit smoking. Split-population hazard models are estimated for both starting and quitting smoking. The analysis suggests price plays a significant role in the decision to start smoking but not in the decision to quit. Further sensitivity analysis of different age groups and an alternative data source, questions the robustness of the significant role of price in the smoking initiation decision. From a policy perspective, the results indicate that increases in tobacco taxation can be an important instrument in reducing the incidence of smoking, but should be combined with other mechanisms such as mandating smoke-free environments and antismoking education. Our results strongly support the targeting of antismoking campaigns towards teenagers.

A conformity test of corporate financial disclosures in Asian and Middle East countries

This paper investigates the financial disclosure practices of corporate annual reports published in Asian countries including Bangladesh, Indonesian, Malaysia and the Middle East countries including Bahrain, Iran, Jordan, Kuwait, Oman, Pakistan, Qatar, Saudi Arabia and Turkey. The purpose of the study is to measure the financial disclosure diversity in these countries, with a view to developing a classification of their similarities and differences in respect to their compliance with International Accounting Standards (IASs). Annual reports of 132 public companies listed on the relevant countries stock exchanges are the central data source, supplemented with other reliable information about financial disclosure practices in each country. A disclosure checklist adopted from all IASs and summarised in 306 individual items of financial disclosures is used as a means of extending an understanding of financial reporting literature. Additionally, it provides an indication of voluntary progress towards harmonisation of international accounting practices in several Asian and Middle Eastern countries. Results show the degree of conformity with IASs from less to high conformity is: Saudi Arabia, Qatar, Bangladesh, Iran, Bahrain, Jordan, Pakistan, Oman, Turkey, Malaysia, Kuwait and Indonesia.

Sporting club supporter identification and supporter purchasing behaviour for sponsors' products

This paper examines the actual purchasing behaviour of sporting club supporters of sponsors' products. The data source for this paper was a syndicated study conducted nationally by a large marketing research company on behalf of a sport's competition's governing body and its 16 constituent clubs. This empirical paper examines the usefulness of such a study in terms of its ability to relate product and brand preference to actual purchase decisions, especially in the context of an individual club's sponsorship. Club supporters are compared with the supporters of all Clubs, in order to ascertain differences in purchase behaviour. The findings suggest that customised research is likely to be of greater value to individual clubs, once the benefits of initial, aggregate studies have been exploited.

A scalable and robust framework for agent-based heterogeneous database operation

How to operate database efficiently and unfailingly in agent-based heterogeneous data source environment is becoming a big problem. In this paper, we contribute a framework and develop a couple of agent-oriented matchmakers with logical ring organization structure to match task agents' requests with middleware agents of databases. The middleware agent is a wrapper of a specific database and is run on the same server with the database management system. The matchmaker is of the features of proliferation and self-cancellation according to the sensory input from its environment. The ring-based coordination mechanism of matchmakers is designed. Two kinds of matchmakers, namely, host and duplicate, are designed for improving the scalability and robustness of agent-based system. The middleware agents are improved for satisfying the framework. We demonstrate the potentials of the framework by case study and present theoretical and empirical evidence that our approach is scalable and robust.

Association between obesity and diabetes using nested linear models

A data source with 7260 cases from the prospective observational study was used. Cases included both sexes and age ranging between 20yr and 7Oyr. Chi-squared tests and linear models were used to examine the effects of age (linear and quadratic), sex, smoking habit (currently a smoker or non-smoker), presence or absence of type 2 diabetes on obesity as measured by body mass index (BMI) and waist to hip ratio (WHR). Both these measures of
obesity were significantly affected by age, sex and the presence of diabetes. Cases with diabetes were significantly more obese (10%) between the ages 40yr to 59yr compared to patients without diabetes. These results accentuate the need to further examine the association between age, diabetes and obesity and other risk factors with cardiovascular diseases.

Medical error and decision making : learning from the past and present in intensive care

Background : Human error occurs in every occupation. Medical errors may result in a near miss or an actual injury to a patient that has nothing to do with the underlying medical condition. Intensive care has one of the highest incidences of medical error and patient injury in any specialty medical area; thought to be related to the rapidly changing patient status and complex diagnoses and treatments.

Purpose : The aims of this paper are to: (1) outline the definition, classifications and aetiology of medical error; (2) summarise key findings from the literature with a specific focus on errors arising from intensive care areas; and (3) conclude with an outline of approaches for analysing clinical information to determine adverse events and inform practice change in intensive care.

Data source : Database searches of articles and textbooks using keywords: medical error, patient safety, decision making and intensive care. Sociology and psychology literature cited therein.

Findings : Critically ill patients require numerous medications, multiple infusions and procedures. Although medical errors are often detected by clinicians at the bedside, organisational processes and systems may contribute to the problem. A systems approach is thought to provide greater insight into the contributory factors and potential solutions to avoid preventable adverse events.

Conclusion : It is recommended that a variety of clinical information and research techniques are used as a priority to prevent hospital acquired injuries and address patient safety concerns in intensive care.

An Iranian migrant mother's production of selves through her family photograph collection

In researching migrant identities, visual methodology offers much promise and yet there is a marked lack of recent research that makes use of visual methods. Previous studies of migrant identities have privileged verbal representations of identities. Gillian Rose's (2010) work with mothers and their family photographs in the United Kingdom, in which she describes the role of family photographs in both producing social subject positions and in maintaining family togetherness across distances, is a useful model for research into the construction and negotiation of migrant identities.

The literature on family photography suggests that it is usually the woman/mother who takes on the role of making the family photograph album; of narrating the family's story (Rose, 2010; Holland, 2009; 1991; Chambers, 2003). The family photograph collection, together with the participant's interwoven verbal interpretation, is a particularly relevant data source for use in identity research as there is the potential for key themes of place, mobilities and space to be explored at new depth and from a feminist perspective.

This paper will report on an ethics approved study of one Iranian migrant mother and her family photograph collection, focusing on her representation of the identities and subjectivities of herself and her children through photographs taken following migration to Australia. The paper will consider the participant's family photographs as both visual objects and visual-verbal narratives produced within historically and culturally situated discourses. It will explore how photographs and oral interpretations cohere to enable migrant mothers to re/produce selves. The paper will examine the production of subject positions specifically in relation to place, mobilities and space.

The research is situated within critical visual ethnography and is informed by a reflexive feminist approach. The meaning making of the photographs under study will be explored at the sites of production, image, and audience. The combined visual-verbal methods used in this study aim to provide a new contribution to the literature on migrant identities and form the basis of a scaled up research design of a larger cohort of mothers.

Do we know how to design effective health coaching interventions: a systematic review of the state of the literature

 Objective. To systematically review health coaching interventions regarding effectiveness of health coaching for specific outcomes, optimal intervention approaches, and identification of specific techniques associated with effectiveness.
Data Source. Articles were sourced from CINAHL, Global Health, PsycINFO, Academic Search Complete, Health Source, Psychology and Behavioral Sciences Collection, and Medline.
Study Inclusion and Exclusion Criteria. Randomized controlled trials were included if the study (1) employed health coaching according to a predefined criterion; (2) clearly reported the use of health coaching; or (3) incorporated the use of coaching.

Data Extraction. Aims, participants, approach, behavior change techniques (BCTs), and findings pertaining to each study were summarized. BCTs were classified according to the CALO-RE taxonomy.

Data Synthesis. Data were synthesized by cross-tabulation of BCTs with study outcomes.
Results. Fifteen of 16 eligible studies reported a positive intervention effect in at least one outcome. Nine studies (56%) did not define health coaching; the number of intervention sessions provided ranged from 2 to 48; and in three studies, one or more intervention details were unclear. It was hence difficult to synthesize the studies to adequately address our research questions.
Conclusion. Health coaching is a promising strategy for health improvements; however, future research should ensure clarity in reporting intervention details, clearer definitions of health coaching/theoretical bases, consistency in reporting BCTs, and the inclusion of process variables as outcome measures.

Perceived fairness of pay among people with and without disabilities: a propensity score matched analysis of working Australians

OBJECTIVES: Equity and fairness at work are associated with a range of organizational and health outcomes. Past research suggests that workers with disabilities experience inequity in the workplace. It is difficult to conclude whether the presence of disability is the reason for perceived unfair treatment due to the possible confounding of effect estimates by other demographic or socioeconomic factors. METHODS: The data source was the Household, Income, and Labor Dynamics in Australia (HILDA) survey (2001-2012). Propensity for disability was calculated from logistic models including gender, age, education, country of birth, and father's occupational skill level as predictors. We then used nearest neighbor (on propensity score) matched analysis to match workers with disabilities to workers without disability. RESULTS: Results suggest that disability is independently associated with lower fairness of pay after controlling for confounding factors in the propensity score matched analysis; although results do suggest less than half a standard deviation difference, indicating small effects. Similar results were apparent in standard multivariable regression models and alternative propensity score analyses (stratification, covariate adjustment using the propensity score, and inverse probability of treatment weighting). CONCLUSIONS: Whilst neither multivariable regression nor propensity scores adjust for unmeasured confounding, and there remains the potential for other biases, similar results for the two methodological approaches to confounder adjustment provide some confidence of an independent association of disability with perceived unfairness of pay. Based on this, we suggest that the disparity in the perceived fairness of pay between people with and without disabilities may be explained by worse treatment of people with disabilities in the workplace.

Noncommunicable disease prevention and control in Mongolia: a policy analysis

BACKGROUND: Noncommunicable diseases (NCDs) are the major global cause of morbidity and mortality. In Mongolia, a number of health policies have been developed targeting the prevention and control of noncommunicable diseases. This paper aimed to evaluate the extent to which NCD-related policies introduced in Mongolia align with the World Health Organization (WHO) 2008-2013 Action Plan for the Global Strategy for the Prevention and Control of NCDs. METHODS: We conducted a review of policy documents introduced by the Government of Mongolia from 2000 to 2013. A literature review, internet-based search, and expert consultation identified the policy documents. Information was extracted from the documents using a matrix, mapping each document against the six objectives of the WHO 2008-2013 Action Plan for the Global Strategy for the Prevention and Control of NCDs and five dimensions: data source, aim and objectives of document, coverage of conditions, coverage of risk factors and implementation plan. 45 NCD-related policies were identified. RESULTS: Prevention and control of the common NCDs and their major risk factors as described by WHO were widely addressed, and policies aligned well with the objectives of the WHO 2008-2013 Action Plan for the Global Strategy for the Prevention and Control of NCDs. Many documents included explicit implementation or monitoring frameworks. It appears that each objective of the WHO 2008-2013 NCD Action Plan was well addressed. Specific areas less well and/or not addressed were chronic respiratory disease, physical activity guidelines and dietary standards. CONCLUSIONS: The Mongolian Government response to the emerging burden of NCDs is a population-based public health approach that includes a national multisectoral framework and integration of NCD prevention and control policies into national health policies. Our findings suggest gaps in addressing chronic respiratory disease, physical activity guidelines, specific food policy actions restricting sales advertising of food products, and a lack of funding specifically supporting NCD research. The neglect of these areas may hamper addressing the NCD burden, and needs immediate action. Future research should explore the effectiveness of national NCD policies and the extent to which the policies are implemented in practice.