Inference in credal networks using multilinear programming

A credal network is a graphical tool for representation and manipulation of uncertainty, where probability values may be imprecise or indeterminate. A credal network associates a directed acyclic graph with a collection of sets of probability measures; in this context, inference is the computation of tight lower and upper bounds for conditional probabilities. In this paper we present new algorithms for inference in credal networks based on multilinear programming techniques. Experiments indicate that these new algorithms have better performance than existing ones, in the sense that they can produce more accurate results in larger networks.

Validation of non‑stationary precipitation series for site‑specific impact assessment: comparison of two statistical downscaling techniques

Statistical downscaling (SD) methods have become a popular, low-cost and accessible means of bridging the gap between the coarse spatial resolution at which climate models output climate scenarios and the finer spatial scale at which impact modellers require these scenarios, with various different SD techniques used for a wide range of applications across the world. This paper compares the Generator for Point Climate Change (GPCC) model and the Statistical DownScaling Model (SDSM)—two contrasting SD methods—in terms of their ability to generate precipitation series under non-stationary conditions across ten contrasting global climates. The mean, maximum and a selection of distribution statistics as well as the cumulative frequencies of dry and wet spells for four different temporal resolutions were compared between the models and the observed series for a validation period. Results indicate that both methods can generate daily precipitation series that generally closely mirror observed series for a wide range of non-stationary climates. However, GPCC tends to overestimate higher precipitation amounts, whilst SDSM tends to underestimate these. This infers that GPCC is more likely to overestimate the effects of precipitation on a given impact sector, whilst SDSM is likely to underestimate the effects. GPCC performs better than SDSM in reproducing wet and dry day frequency, which is a key advantage for many impact sectors. Overall, the mixed performance of the two methods illustrates the importance of users performing a thorough validation in order to determine the influence of simulated precipitation on their chosen impact sector.

A new statistical framework for the quantification of covariate associations with species distributions

Identifying processes that shape species geographical ranges is a prerequisite for understanding environmental change. Currently, species distribution modelling methods do not offer credible statistical tests of the relative influence of climate factors and typically ignore other processes (e.g. biotic interactions and dispersal limitation). We use a hierarchical model fitted with Markov Chain Monte Carlo to combine ecologically plausible niche structures using regression splines to describe unimodal but potentially skewed response terms. We apply spatially explicit error terms that account for (and may help identify) missing variables. Using three example distributions of European bird species, we map model results to show sensitivity to change in each covariate. We show that the overall strength of climatic association differs between species and that each species has considerable spatial variation in both the strength of the climatic association and the sensitivity to climate change. Our methods are widely applicable to many species distribution modelling problems and enable accurate assessment of the statistical importance of biotic and abiotic influences on distributions.

The subjective numeracy scale as a tool for assessing statistical numeracy in older adult populations

BACKGROUND:
Statistical numeracy, necessary for making informed medical decisions, is reduced among older adults who make more decisions about their medical care and treatment than at any other stage of life. Objective numeracy scales are a source of anxiety among patients, heightened among older adults.
OBJECTIVE:
We investigate the subjective numeracy scale as an alternative tool for measuring statistical numeracy with older adult samples.
METHODS:
Numeracy was assessed using objective measures for 526 adults ranging in age from 18 to 93 years, and all participants provided subjective numeracy ratings.
RESULTS:
Subjective numeracy correlated highly with objective measurements among oldest adults (70+ years; r = 0.51, 95% CI 0.32, 0.66), and for younger age groups. Subjective numeracy explained 33.2% of age differences in objective numeracy.
CONCLUSION:
The subjective numeracy scale provides an effective tool for assessing statistical numeracy for broad age ranges and circumvents problems associated with objective numeracy measures.

Statistical Analysis of Small Ellerman Bomb Events

The properties of Ellerman bombs (EBs), small-scale brightenings in the Hα line wings, have proved difficult to establish because their size is close to the spatial resolution of even the most advanced telescopes. Here, we aim to infer the size and lifetime of EBs using high-resolution data of an emerging active region collected using the Interferometric BIdimensional Spectrometer (IBIS) and Rapid Oscillations of the Solar Atmosphere (ROSA) instruments as well as the Helioseismic and Magnetic Imager (HMI) onboard the Solar Dynamics Observatory (SDO). We develop an algorithm to track EBs through their evolution, finding that EBs can often be much smaller (around 0.3″) and shorter-lived (less than one minute) than previous estimates. A correlation between G-band magnetic bright points and EBs is also found. Combining SDO/HMI and G-band data gives a good proxy of the polarity for the vertical magnetic field. It is found that EBs often occur both over regions of opposite polarity flux and strong unipolar fields, possibly hinting at magnetic reconnection as a driver of these events.The energetics of EB events is found to follow a power-law distribution in the range of a nanoflare (1022-25 ergs).

A survey of location inference techniques on Twitter

The increasing popularity of the social networking service, Twitter, has made it more involved in day-to-day communications, strengthening social relationships and information dissemination. Conversations on Twitter are now being explored as indicators within early warning systems to alert of imminent natural disasters such earthquakes and aid prompt emergency responses to crime. Producers are privileged to have limitless access to market perception from consumer comments on social media and microblogs. Targeted advertising can be made more effective based on user profile information such as demography, interests and location. While these applications have proven beneficial, the ability to effectively infer the location of Twitter users has even more immense value. However, accurately identifying where a message originated from or author’s location remains a challenge thus essentially driving research in that regard. In this paper, we survey a range of techniques applied to infer the location of Twitter users from inception to state-of-the-art. We find significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.

Using interviewer random effects to remove selection bias from HIV prevalence estimates

Background: Selection bias in HIV prevalence estimates occurs if non-participation in testing is correlated with HIV status. Longitudinal data suggests that individuals who know or suspect they are HIV positive are less likely to participate in testing in HIV surveys, in which case methods to correct for missing data which are based on imputation and observed characteristics will produce biased results. Methods: The identity of the HIV survey interviewer is typically associated with HIV testing participation, but is unlikely to be correlated with HIV status. Interviewer identity can thus be used as a selection variable allowing estimation of Heckman-type selection models. These models produce asymptotically unbiased HIV prevalence estimates, even when non-participation is correlated with unobserved characteristics, such as knowledge of HIV status. We introduce a new random effects method to these selection models which overcomes non-convergence caused by collinearity, small sample bias, and incorrect inference in existing approaches. Our method is easy to implement in standard statistical software, and allows the construction of bootstrapped standard errors which adjust for the fact that the relationship between testing and HIV status is uncertain and needs to be estimated. Results: Using nationally representative data from the Demographic and Health Surveys, we illustrate our approach with new point estimates and confidence intervals (CI) for HIV prevalence among men in Ghana (2003) and Zambia (2007). In Ghana, we find little evidence of selection bias as our selection model gives an HIV prevalence estimate of 1.4% (95% CI 1.2% – 1.6%), compared to 1.6% among those with a valid HIV test. In Zambia, our selection model gives an HIV prevalence estimate of 16.3% (95% CI 11.0% - 18.4%), compared to 12.1% among those with a valid HIV test. Therefore, those who decline to test in Zambia are found to be more likely to be HIV positive. Conclusions: Our approach corrects for selection bias in HIV prevalence estimates, is possible to implement even when HIV prevalence or non-participation is very high or very low, and provides a practical solution to account for both sampling and parameter uncertainty in the estimation of confidence intervals. The wide confidence intervals estimated in an example with high HIV prevalence indicate that it is difficult to correct statistically for the bias that may occur when a large proportion of people refuse to test.