A Bayesian approach to time-varying latent strengths in pairwise comparisons.

The famous Bradley-Terry model for pairwise comparisons is widely used for ranking objects and is often applied to sports data. In this paper we extend the Bradley-Terry model by allowing time-varying latent strengths of compared objects. The time component is modelled with barycentric rational interpolation and Gaussian processes. We also allow for the inclusion of additional information in the form of outcome probabilities. Our models are evaluated and compared on toy data set and real sports data from ATP tennis matches and NBA games. We demonstrated that using Gaussian processes is advantageous compared to barycentric rational interpolation as they are more flexible to model discontinuities and are less sensitive to initial parameters settings. However, all investigated models proved to be robust to over-fitting and perform well with situations of volatile and of constant latent strengths. When using barycentric rational interpolation it has turned out that applying Bayesian approach gives better results than by using MLE. Performance of the models is further improved by incorporating the outcome probabilities.

Seroprevalence of severe acute respiratory syndrome coronavirus 2 in Slovenia: results of two rounds of a nationwide population study on a probability-based sample, challenges and lessons learned.

OBJECTIVES: Seroprevalence surveys provide crucial information on cumulative severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) exposure. This Slovenian nationwide population study is the first longitudinal 6-month serosurvey using probability-based samples across all age categories. METHODS: Each participant supplied two blood samples: 1316 samples in April 2020 (first round) and 1211 in October/November 2020 (second round). The first-round sera were tested using Euroimmun Anti-SARS-CoV-2 ELISA IgG (ELISA) and, because of uncertain estimates, were retested using Elecsys Anti-SARS-CoV-2 (Elecsys-N) and Elecsys Anti-SARS-CoV-2 S (Elecsys-S). The second-round sera were concomitantly tested using Elecsys-N/Elecsys-S. RESULTS: The populations of both rounds matched the overall population (n = 3000), with minor settlement type and age differences. The first-round seroprevalence corrected for the ELISA manufacturer's specificity was 2.78% (95% highest density interval [HDI] 1.81%-3.80%), corrected using pooled ELISA specificity calculated from published data 0.93% (95% CI 0.00%-2.65%), and based on Elecsys-N/Elecsys-S results 0.87% (95% HDI 0.40%-1.38%). The second-round unadjusted lower limit of seroprevalence on 11 November 2020 was 4.06% (95% HDI 2.97%-5.16%) and on 3 October 2020, unadjusted upper limit was 4.29% (95% HDI 3.18%-5.47%). CONCLUSIONS: SARS-CoV-2 seroprevalence in Slovenia increased four-fold from late April to October/November 2020, mainly due to a devastating second wave. Significant logistic/methodological challenges accompanied both rounds. The main lessons learned were a need for caution when relying on manufacturer-generated assay evaluation data, the importance of multiple manufacturer-independent assay performance assessments, the need for concomitant use of highly-specific serological assays targeting different SARS-CoV-2 proteins in serosurveys conducted in low-prevalence settings or during epidemic exponential growth and the usefulness of a Bayesian approach for overcoming complex methodological challenges.

bayes4psy-An Open Source R Package for Bayesian Statistics in Psychology.

Research in psychology generates complex data and often requires unique statistical analyses. These tasks are often very specific, so appropriate statistical models and methods cannot be found in accessible Bayesian tools. As a result, the use of Bayesian methods is limited to researchers and students that have the technical and statistical fundamentals that are required for probabilistic programming. Such knowledge is not part of the typical psychology curriculum and is a difficult obstacle for psychology students and researchers to overcome. The goal of the bayes4psy package is to bridge this gap and offer a collection of models and methods to be used for analysing data that arises from psychological experiments and as a teaching tool for Bayesian statistics in psychology. The package contains the Bayesian t-test and bootstrapping along with models for analysing reaction times, success rates, and tasks utilizing colors as a response. It also provides the diagnostic, analytic and visualization tools for the modern Bayesian data analysis workflow.

Trends in NBA and Euroleague basketball: Analysis and comparison of statistical data from 2000 to 2017.

We analyse and compare NBA and Euroleague basketball through box-score statistics in the period from 2000 to 2017. Overall, the quantitative differences between the NBA and Euroleague have decreased and are still decreasing. Differences are even smaller after we adjust for game length and when playoff NBA basketball is considered instead of regular season basketball. The differences in factors that contribute to success are also very small-(Oliver's) four factors derived from box-score statistics explain most of the variability in team success even if the coefficients are determined for both competitions simultaneously instead of each competition separately. The largest difference is game pace-in the NBA there are more possessions per game. The number of blocks, the defensive rebounding rate and the number of free throws per foul committed are also higher in the NBA, while the number of fouls committed is lower. Most of the differences that persist can be reasonably explained by the contrasts between the better athleticism of NBA players and more emphasis on tactical aspects of basketball in the Euroleague.

Impact of changes in climate on air pollution in Slovenia between 2002 and 2017.

Air pollutant levels depend on emissions but can also be affected by the meteorological situation. We examined air pollutant trends (PM10, NO2, O3 and SO2) in Slovenia, where in the past the main issue were SO2 levels. Now, the population is still exposed to PM10 and ozone levels that are above the recommended levels. Our goal was to assess if the levels of air pollutants were decreasing from 2002 to 2017 due to emission ceilings or were more influenced by changes in the meteorological situation. We modelled the relationship between levels, meteorological parameters, and seasonality and then used the models with the best estimated generalisation to adjust levels for meteorology. Models showed a significant relationship between meteorological parameters and PM10, NO2, and O3 levels, but not SO2. We analysed trends of raw and adjusted levels and compared them. Trends of PM10 and SO2 were decreasing at all locations for raw and adjusted data. The largest decrease was observed in SO2 levels where the largest decrease in emissions occurred. Trends of NO2 were also significant and negative at most locations. Levels of O3 did not exhibit a significant trend at most locations. Results show that changes in the meteorological situation affected PM10 levels the most, especially where the entire period (2002-2017) could be observed. There is strong empirical evidence that changes in meteorological parameters contributed to the decrease in PM10 levels while the decrease in NO2 and SO2 levels can be attributed to emission ceilings.

A Bayesian hierarchical latent trait model for estimating rater bias and reliability in large-scale performance assessment.

We propose a novel approach to modelling rater effects in scoring-based assessment. The approach is based on a Bayesian hierarchical model and simulations from the posterior distribution. We apply it to large-scale essay assessment data over a period of 5 years. Empirical results suggest that the model provides a good fit for both the total scores and when applied to individual rubrics. We estimate the median impact of rater effects on the final grade to be ± 2 points on a 50 point scale, while 10% of essays would receive a score at least ± 5 different from their actual quality. Most of the impact is due to rater unreliability, not rater bias.

Bayesian Lasso and multinomial logistic regression on GPU.

We describe an efficient Bayesian parallel GPU implementation of two classic statistical models-the Lasso and multinomial logistic regression. We focus on parallelizing the key components: matrix multiplication, matrix inversion, and sampling from the full conditionals. Our GPU implementations of Bayesian Lasso and multinomial logistic regression achieve 100-fold speedups on mid-level and high-end GPUs. Substantial speedups of 25 fold can also be achieved on older and lower end GPUs. Samplers are implemented in OpenCL and can be used on any type of GPU and other types of computational units, thereby being convenient and advantageous in practice compared to related work.

A Balanced Mixture of Antagonistic Pressures Promotes the Evolution of Parallel Movement.

A common hypothesis about the origins of collective behaviour suggests that animals might live and move in groups to increase their chances of surviving predator attacks. This hypothesis is supported by several studies that use computational models to simulate natural evolution. These studies, however, either tune an ad-hoc model to 'reproduce' collective behaviour, or concentrate on a single type of predation pressure, or infer the emergence of collective behaviour from an increase in prey density. In nature, prey are often targeted by multiple predator species simultaneously and this might have played a pivotal role in the evolution of collective behaviour. We expand on previous research by using an evolutionary rule-based system to simulate the evolution of prey behaviour when prey are subject to multiple simultaneous predation pressures. We analyse the evolved behaviour via prey density, polarization, and angular momentum. Our results suggest that a mixture of antagonistic external pressures that simultaneously steer prey towards grouping and dispersing might be required for prey individuals to evolve dynamic parallel movement.

Predicting the outcome of head-up tilt test using heart rate variability and baroreflex sensitivity parameters in patients with vasovagal syncope.

PURPOSE: The aim of the study was to investigate whether a statistical model could be used for an early prediction of the head-up tilt test (HUTT) outcome from heart rate variability (HRV) and baroreflex sensitivity (BRS) data obtained during early stages of the HUTT. METHODS: A modified Italian protocol was used for HUTT in 105 patients with a previous history of vasovagal syncope. Beat-to-beat heart rate and blood pressure were continuously recorded. Fast Fourier transformation was used for spectral analysis of HRV and a sequence technique for measuring the BRS. RESULTS: Linear statistical models based on HRV and BRS data from the first 15 min of HUTT were no more accurate than always naively predicted majority class that a syncope will occur (average model out-of-sample accuracy 56.2 ± 5.1 % vs. majority class relative frequency 54.2 %). Even when HRV and BRS data from the first 30 min were used in the model, we did not obtain any predictions of meaningful practical value (75.0 ± 5.1 % accuracy vs. 72.2 % majority class). CONCLUSIONS: While there are discernible and meaningful differences between HUTT-P and HUTT-N subjects, they are not sufficient to discriminate between the two groups and predict a syncope early in the HUTT. The results might improve with a larger set of subjects; however, we can conclude that it is not likely that syncope predictions of practical value can be obtained from aggregate HRV spectral analysis and BRS values.

Basketball shot types and shot success in different levels of competitive basketball.

The purpose of our research was to investigate the relative frequencies of different types of basketball shots (above head, hook shot, layup, dunk, tip-in), some details about their technical execution (one-legged, two-legged, drive, cut, …), and shot success in different levels of basketball competitions. We analysed video footage and categorized 5024 basketball shots from 40 basketball games and 5 different levels of competitive basketball (National Basketball Association (NBA), Euroleague, Slovenian 1st Division, and two Youth basketball competitions). Statistical analysis with hierarchical multinomial logistic regression models reveals that there are substantial differences between competitions. However, most differences decrease or disappear entirely after we adjust for differences in situations that arise in different competitions (shot location, player type, and attacks in transition). Differences after adjustment are mostly between the Senior and Youth competitions: more shots executed jumping or standing on one leg, more uncategorised shot types, and more dribbling or cutting to the basket in the Youth competitions, which can all be attributed to lesser technical and physical ability of developing basketball players. The two discernible differences within the Senior competitions are that, in the NBA, dunks are more frequent and hook shots are less frequent compared to European basketball, which can be attributed to better athleticism of NBA players. The effect situational variables have on shot types and shot success are found to be very similar for all competitions.

Analysis of experts' quantitative assessment of adolescent basketball players and the role of anthropometric and physiological attributes.

IN THIS PAPER, WE INVESTIGATED TWO QUESTIONS: (1) can measurements of anthropometric and physiological attributes substitute for expert assessment of adolescent basketball players, and (2) how much does the quantitative assessment of a player vary among experts? The first question is relevant to the potential simplification of the player selection process. The second question pertains directly to the validity of expert quantitative assessment. Our research was based on data from 148 U14 female and male basketball players. For each player, an array of anthropometric and physiological attributes was recorded, including body height, body mass, BMI, and several motor skill tests. Furthermore, each player's current ability and potential ability were quantitatively evaluated by two different experts from a group of seven experts. Analysis of the recorded data showed that the anthropometric and physiological attributes explained between 15% and 40% of the variance in experts' scores. The primary predictive attributes were speed and agility (for predicting current ability) and body height and growth potential (for predicting potential ability). We concluded that these attributes were not sufficiently informative to act as a substitute for expert assessment of the players' current or potential ability. There is substantial variability in different experts' scores of the same player's ability. However, the differences between experts are mostly in scale, and the relationships between experts' scores are monotonic. That is, different experts rank players on ability very similarly, but their scores are not well calibrated.

A decade of euroleague basketball: an analysis of trends and recent rule change effects.

The International Basketball Federation (FIBA) recently introduced major rule changes that came into effect with the 2010/11 season. Most notably, moving the three-point arc and changing the shot-clock. The purpose of this study was to investigate and quantify how these changes affect the game performance of top-level European basketball players. In order to better understand these changes, we also investigated past seasons and showed the presence of several trends, even in the absence of significant rule changes. A large set of game statistics for 10 seasons and 2198 Euroleague basketball games in which top European clubs competed was analyzed. Results show that the effects of the rule changes are contrary to trends in recent years.