Using data mining to predict road crash count with a focus on skid resistance values

Road crashes cost world and Australian society a significant proportion of GDP, affecting productivity and causing significant suffering for communities and individuals. This paper presents a case study that generates data mining models that contribute to understanding of road crashes by allowing examination of the role of skid resistance (F60) and other road attributes in road crashes. Predictive data mining algorithms, primarily regression trees, were used to produce road segment crash count models from the road and traffic attributes of crash scenarios. The rules derived from the regression trees provide evidence of the significance of road attributes in contributing to crash, with a focus on the evaluation of skid resistance.

History behind the egg: Women's comprehensive histories and IVF outcomes, Australia, 2008-2009 [abstract]

The individual history of infertile women, as well as their age, may influence their response to in vitro fertilisation (IVF) cycles. This study examined the associations between women’s histories and two IVF outcomes: eggs aspirated (EA) and proportion with normal, two-pronuclei (2PN), fertilisation. This is a cross-sectional survey of infertile women (n=141, 27-46 years) from a multi-centre clinical sample. Participants completed a survey of socio-demographic, relationship, lifestyle, reproductive and fertility factors, medical conditions and recurrent symptoms. Among participants with heterosexual partners (n=122), associations between women’s histories and EA or 2PN fertilisation were analysed using linear and logistic modelling, respectively, adjusted for age at EA and accounting for multiple IVF cycles (n=313 cycles). Participants aged 35+ years had reproductive histories of miscarriage only (16.9%), termination only (9.9%) or birth+termination (5.6%) that were 2-, 3- and 4-fold higher, respectively, than those aged <35 years (7.1%, 2.9%, 1.4%). More years of oral contraceptive use were associated with a lower mean EA: never used, 14.6 EA; 0-2 years, 11.7 EA; 3-5 years, 8.6 EA; 6þ years, 8.2 EA (p=.04). Participants with polycystic ovary syndrome had a higher mean EA (11.5) than those without the condition (8.3 EA, p<.01). Participants in trade or service occupations had lower proportions of 2PN fertilisation (51.7%) than participants in other occupations (professional, 58.6%; manual/other, 63.6%, p<.02). Increasing women’s age and prolonged used of oral contraceptives were associated with lower EA from IVF cycles; PCOS was associated with higher EA. Occupational exposures may have a detrimental effect on normal fertilisation rates.

History behind the IVF egg: Are women's comprehensive histories associated with the number of eggs collected, or fertilised normally, from repeated IVF cycles? [abstract]

Most studies of in vitro fertilisation (IVF) outcomes use cycle-based data and fail to account for women who use repeated IVF cycles. The objective of this study was to examine the association between the number of eggs collected (EC) and the percentage fertilised normally, and women’s self-reported medical, personal and social histories. This study involved a crosssectional survey of infertile women (aged 27-46 years) recruited from four privately-owned fertility clinics located in major cities of Australia. Regression modeling was used to estimate the mean EC and mean percentage of eggs fertilised normally: adjusted for age at EC. Appropriate statistical methods were used to take account of repeated IVF cycles by the same women. Among 121 participants who returned the survey and completed 286 IVF cycles, the mean age at EC was 35.2 years (SD 4.5). Women’s age at EC was strongly associated with the number of EC: <30 years, 11.7 EC; 30.0-< 35 years, 10.6 EC; 35.0-<40.0 years, 7.3 EC; 40.0+ years, 8.1 EC; p<.0001. Prolonged use of oral contraceptives was associated with lower numbers of EC: never used, 14.6 EC; 0-2 years, 11.7 EC; 3-5 years, 8.5 EC; 6þ years, 8.2 EC; p=.04. Polycystic ovary syndrome (PCOS) was associated with more EC: have PCOS, 11.5 EC; no, 8.3 EC; p=.01. Occupational exposures may be detrimental to normal fertilisation: professional roles, 58.8%; trade and service roles, 51.8%; manual and other roles, 63.3%; p=.02. In conclusion, women’s age remains the most significant characteristic associated with EC but not the percentage of eggs fertilised normally.

When the right to be counted doesn’t count : the politics and challenges of researching the health of asylum seekers

A fundamental prerequisite of population health research is the ability to establish an accurate denominator. This in turn requires that every individual in the study population is counted. However, this seemingly simple principle has become a point of conflict between researchers whose aim is to produce evidence of disparities in population health outcomes and governments whose policies promote(intentionally or not) inequalities that are the underlying causes of health disparities. Research into the health of asylum seekers is a case in point. There is a growing body of evidence documenting the adverse affects of recent changes in asylum-seeking legislation, including mandatory detention. However, much of this evidence has been dismissed by some governments as being unsound, biased and unscientific because, it is argued, evidence is derived from small samples or from case studies. Yet, it is the policies of governments that are the key barrier to the conduct of rigorous population health research on asylum seekers. In this paper, the authors discuss the challenges of counting asylum seekers and the limitations of data reported in some industrialized countries. They argue that the lack of accurate statistical data on asylum seekers has been an effective neo-conservative strategy for erasing the health inequalities in this vulnerable population, indeed a strategy that renders invisible this population. They describe some alternative strategies that may be used by researchers to obtain denominator data on hard-to-reach populations such as asylum seekers.

Does 'habitus' count in Chinese Australians' mathematics achievement?

Large-scale international comparative studies and cross-ethnic studies have revealed that Chinese students, whether living in China or overseas, consistently outperform their counterparts in mathematics achievement. These studies tended to explain this result from psychological, educational, or cultural perspectives. However, there is scant sociological investigation addressing Chinese students’ better mathematics achievement. Drawing on Bourdieu’s sociological theory, this study conceptualises Chinese Australians’ “Chineseness” by the notion of ‘habitus’ and considers this “Chineseness” generating but not determinating mechanism that underpins Chinese Australians’ mathematics learning. Two hundred and thirty complete responses from Chinese Australian participants were collected by an online questionnaire. Simple regression model statistically significantly well predicted mathematics achievement by “Chineseness” (F = 141.90, R = .62, t = 11.91, p < .001). Taking account of “Chineseness” as a sociological mechanism for Chinese Australians’ mathematics learning, this study complements psychological and educational impacts on better mathematics achievement of Chinese students revealed by previous studies. This study also challenges the cultural superiority discourse that attributes better mathematics achievement of Chinese students to cultural factors.

Exact and approximate Bayesian inference for low count time series models with intractable likelihoods

In this paper we present a new simulation methodology in order to obtain exact or approximate Bayesian inference for models for low-valued count time series data that have computationally demanding likelihood functions. The algorithm fits within the framework of particle Markov chain Monte Carlo (PMCMC) methods. The particle filter requires only model simulations and, in this regard, our approach has connections with approximate Bayesian computation (ABC). However, an advantage of using the PMCMC approach in this setting is that simulated data can be matched with data observed one-at-a-time, rather than attempting to match on the full dataset simultaneously or on a low-dimensional non-sufficient summary statistic, which is common practice in ABC. For low-valued count time series data we find that it is often computationally feasible to match simulated data with observed data exactly. Our particle filter maintains $N$ particles by repeating the simulation until $N+1$ exact matches are obtained. Our algorithm creates an unbiased estimate of the likelihood, resulting in exact posterior inferences when included in an MCMC algorithm. In cases where exact matching is computationally prohibitive, a tolerance is introduced as per ABC. A novel aspect of our approach is that we introduce auxiliary variables into our particle filter so that partially observed and/or non-Markovian models can be accommodated. We demonstrate that Bayesian model choice problems can be easily handled in this framework.

Fecal indicators and bacterial pathogens in bottled water from Dhaka, Bangladesh

Forty-six bottled water samples representing 16 brands from Dhaka, Bangladesh were tested for the numbers of total coliforms, fecal indicator bacteria (i.e., thermotolerant Escherichia coli and Enterococcus spp.) and potential bacterial pathogens (i.e., Aeromonas hydrophil, Pseudomonas aeruginos, Salmonella spp., and Shigella spp.). Among the 16 brands tested, 14 (86%), ten (63%) and seven (44%) were positive for total coliforms, E. coil and Enterococcus spp., respectively. Additionally, a further nine (56%), eight (50%), six (37%), and four (25%) brands were PCR positive for A. hydrophila lip, P. aeruginosa ETA, Salmonella spp. invA, and Shigella spp. ipaH genes, respectively. The numbers of bacterial pathogens in bottled water samples ranged from 28 ± 12 to 600 ± 45 (A. hydrophila lip gene), 180 ± 40 to 900 ± 200 (Salmonella spp. invA gene), 180 ± 40 to 1,300 ± 400 (P. aeruginosa ETA gene) genomic units per L of water. Shigella spp. ipaH gene was not quantifiable. Discrepancies were observed in terms of the occurrence of fecal indicators and bacterial pathogens. No correlations were observed between fecal indicators numbers and presence/absence of A. hydrophila lip (p = 0.245), Salmonella spp. invA (p = 0.433), Shigella spp. ipaH gene (p = 0.078), and P. aeruginosa ETA (p = 0.059) genes. Our results suggest that microbiological quality of bottled waters sold in Dhaka, Bangladesh is highly variable. To protect public health, stringent quality control is recommended for the bottled water industry in Bangladesh. Key words: bottled water, fecal indicator bacteria, quantitative PCR, bacterial pathogens, public health risk.

Count-min sketches for estimating password frequency within Hamming distance two

The count-min sketch is a useful data structure for recording and estimating the frequency of string occurrences, such as passwords, in sub-linear space with high accuracy. However, it cannot be used to draw conclusions on groups of strings that are similar, for example close in Hamming distance. This paper introduces a variant of the count-min sketch which allows for estimating counts within a specified Hamming distance of the queried string. This variant can be used to prevent users from choosing popular passwords, like the original sketch, but it also allows for a more efficient method of analysing password statistics.

Mining universities : can we count on the integrity of research coming out of today's mining-funded tertiary sector?

Australia's economic growth and national identity have been widely celebrated as being founded on the nation's natural resources. With the golden era of pastoralism fading into the distance, a renewed love affair with primary industries has been much lauded, particularly by purveyors of neoliberal ideology. The considerable wealth generated by resource extraction has, despite its environmental and social record, proved seductive to the university sector. The mining industry is one of a number of industries and sectors (alongside pharmaceutical, chemical and biotechnological) that is increasingly courting Australian universities. These new public-private alliances are often viewed as the much-needed cash cow to bridge the public funding shortfall in the tertiary sector. However, this trend also raises profound questions about the capacity of public good institutions, as universities were once assumed to be, to maintain institutional independence and academic freedoms.

Does counting still count? Revisiting the security of counting based user authentication protocols against statistical attacks

At NDSS 2012, Yan et al. analyzed the security of several challenge-response type user authentication protocols against passive observers, and proposed a generic counting based statistical attack to recover the secret of some counting based protocols given a number of observed authentication sessions. Roughly speaking, the attack is based on the fact that secret (pass) objects appear in challenges with a different probability from non-secret (decoy) objects when the responses are taken into account. Although they mentioned that a protocol susceptible to this attack should minimize this difference, they did not give details as to how this can be achieved barring a few suggestions. In this paper, we attempt to fill this gap by generalizing the attack with a much more comprehensive theoretical analysis. Our treatment is more quantitative which enables us to describe a method to theoretically estimate a lower bound on the number of sessions a protocol can be safely used against the attack. Our results include 1) two proposed fixes to make counting protocols practically safe against the attack at the cost of usability, 2) the observation that the attack can be used on non-counting based protocols too as long as challenge generation is contrived, 3) and two main design principles for user authentication protocols which can be considered as extensions of the principles from Yan et al. This detailed theoretical treatment can be used as a guideline during the design of counting based protocols to determine their susceptibility to this attack. The Foxtail protocol, one of the protocols analyzed by Yan et al., is used as a representative to illustrate our theoretical and experimental results.