The Simpson’s Paradox in the Offline Evaluation of Recommendation Systems

Recommendation systems are often evaluated based on user’s interactions that were collected from an existing, already deployed recommendation system. In this situation, users only provide feedback on the exposed items and they may not leave feedback on other items since they have not been exposed to them by the deployed system. As a result, the collected feedback dataset that is used to evaluate a new model is influenced by the deployed system, as a form of closed loop feedback. In this article, we show that the typical offline evaluation of recommender systems suffers from the so-called Simpson’s paradox. Simpson’s paradox is the name given to a phenomenon observed when a significant trend appears in several different sub-populations of observational data but disappears or is even reversed when these sub-populations are combined together. Our in-depth experiments based on stratified sampling reveal that a very small minority of items that are frequently exposed by the deployed system plays a confounding factor in the offline evaluation of recommendation systems. In addition, we propose a novel evaluation methodology that takes into account the confounder, i.e., the deployed system’s characteristics. Using the relative comparison of many recommendation models as in the typical offline evaluation of recommender systems, and based on the Kendall rank correlation coefficient, we show that our proposed evaluation methodology exhibits statistically significant improvements of 14% and 40% on the examined open loop datasets (Yahoo! and Coat), respectively, in reflecting the true ranking of systems with an open loop (randomised) evaluation in comparison to the standard evaluation.

Drivers of Stock Prices in Ghana: An Empirical Mode Decomposition Approach

This study utilized the empirical mode decomposition (EMD) technique and examined which group of investors based on their trading frequencies influence stock prices in Ghana. We applied this technique to a dataset of daily closing prices of GSE Financial Stock Index for the period 04/01/2011 to 28/08/2015. The daily closing prices were decomposed into six intrinsic mode functions (IMFs) and a residue. We used the hierarchical clustering method to reconstruct the IMFs into high frequency, low frequency, and trend components. Using statistical measures such as Pearson product moment correlation coefficient and the Kendall rank correlation, we found that the low frequency and trend components of stock prices are the main drivers of the GSE stock index. These low-frequency traders are the institutional investors. Therefore, stock prices on the GSE are affected by real economic growth but not short-lived market fluctuations.

Regional analysis of multivariate compound coastal flooding potential around Europe and environs: sensitivity analysis and spatial patterns

In coastal regions, floods can arise through a combination of multiple drivers, including direct surface run-off, river discharge, storm surge, and waves. In this study, we analyse compound flood potential in Europe and environs caused by these four main flooding sources using state-of-the-art databases with coherent forcing (i.e. ERA5). First, we analyse the sensitivity of the compound flooding potential to several factors: (1) sampling method, (2) time window to select the concurrent event of the conditioned driver, (3) dependence metrics, and (4) wave-driven sea level definition. We observe higher correlation coefficients using annual maxima than peaks over threshold. Regarding the other factors, our results show similar spatial distributions of the compound flooding potential. Second, the dependence between the pairs of drivers using the Kendall rank correlation coefficient and the joint occurrence are synthesized for coherent patterns of compound flooding potential using a clustering technique. This quantitative multi-driver assessment not only distinguishes where overall compound flooding potential is the highest, but also discriminates which driver combinations are more likely to contribute to compound flooding. We identify that hotspots of compound flooding potential are located along the southern coast of the North Atlantic Ocean and the northern coast of the Mediterranean Sea.

The development of professionally important qualities as a result of various activities: Learning, sport, and labour market (El desarrollo de cualidades profesionales importantes como resultado de diversas actividades: Aprendizaje, deporte y mercado labo

The purpose of this article is to define evolution of opinions about the formation of the in-demand qualities (soft skills) in the Ukrainian labour market and the possibilities of their formation through sports based on the opinions of key participants in the market. The basis of the study is a survey of professionals (104 respondents), students (314 respondents), and schoolchildren (326 respondents). The coherence of the respondents' answers was checked by the Spearman rank correlation coefficient and Kendall rank correlation coefficient. The correlation, factor, and cluster analyses were used in the study methods. Factor analysis allowed us to form obtained qualities into two factors (of primary and secondary importance). Using cluster analysis, all the qualities that were identified by the respondents were divided into three uniform groups. It was also proved that the clusters differ most in terms of the qualities that were named by schoolchildren, and the clusters which are the closest in terms of the qualities were highlighted by the employers. The study was conducted within the framework of new approaches to searching the tendencies and key elements of competition on the labour markets, and it complements the work (Cachón-Zagalaz et al., 2020) by including the aspects into the analysis of schoolchildren and highlighting the distinguishing aspects that are involved in their perception of in-demand qualities. The uniqueness of the qualities that were highlighted by the students who do sport and the importance of these qualities in the formation of professional competencies have been proved. Resumen. El propósito de este artículo es definir la evolución de las opiniones sobre la formación de las cualidades más demandadas (soft skills) en el mercado laboral ucraniano y las posibilidades de su formación a través del deporte a partir de las opiniones de los principales agentes del mercado. La base del estudio es una encuesta realizada a profesionales (104 encuestados), estudiantes (314 encuestados) y escolares (326 encuestados). La coherencia de las respuestas de los encuestados se comprobó mediante el coeficiente de correlación de rangos de Spearman y el coeficiente de correlación de rangos de Kendall. En los métodos de estudio se utilizaron los análisis de correlación, de factores y de conglomerados. El análisis factorial permitió formar las cualidades obtenidas en dos factores (de importancia primaria y secundaria). Mediante el análisis de conglomerados, todas las cualidades identificadas por los encuestados se dividieron en tres grupos uniformes. También se comprobó que los clusters son los que más se diferencian en cuanto a las cualidades nombradas por los escolares, y los clusters que más se acercan en cuanto a las cualidades fueron destacados por los empresarios. El estudio se ha realizado en el marco de los nuevos enfoques de búsqueda de las tendencias y elementos clave de la competencia en los mercados de trabajo, y complementa el trabajo (Cachón-Zagalaz, et al., 2020) al incluir los aspectos en el análisis de los escolares y destacar los aspectos distintivos que intervienen en su percepción de las cualidades demandadas. Se ha comprobado la singularidad de las cualidades destacadas por los alumnos que hacen deporte y la importancia de las mismas en la formación de competencias profesionales.

Precision Nutrition and Reliability of Continuous Glucose Monitors: Insights From the PREDICT Study

Objectives The use of continuous glucose monitors (CGM) provides a more in-depth characterization of glycemic variation in response to environmental stimuli, but concerns about CGM reliability for categorizing glycemic responses to foods and meals exist. We sought to evaluate the concordance and reliability of two simultaneously worn CGM devices on postprandial glycemic responses. Methods We examined the correlation and coefficient of variation of the 2h glucose incremental area under the curve (iAUC) for 21,527 standardized and ad libitum meals consumed by 368 healthy participants from the PREDICT-1 Study. Included participants were simultaneously monitored with either two Abbott Freestyle Libre Pro devices (n = 339, same device group) or the combination of Abbott Freestyle Libre Pro and Dexcom G6 devices (n = 29, inter device group). Within-subject 2 h iAUC glucose meal rankings for paired CGM devices were assessed using the Kendal-tau measure for ranking concordance. Results The correlation coefficient of the 2 h glucose iAUC for paired CGM devices was 0.97 (95% CI, 0.96 to 0.97) for same device comparisons and 0.78 (0.76 to 0.80) for inter device comparisons. The coefficient of variation of the 2 h glucose iAUC for standardized meals was 5.1% (interquartile range, 2.2 to 10.1) for same device comparisons and 15.1% (5.9 to 31.2) for inter device comparisons. Similar results were observed for ad libitum meals with same and inter device coefficients of variation of 8.9% (3.3 to 21.3) and 24.2% (10.2 to 53.1%), respectively. Meal rankings for the 2 h glucose iAUC were concordant between paired CGM devices, with a mean Kendall rank correlation coefficient of 0.86 (sd = 0.07) for same device comparisons and 0.63 (sd = 0.011) for inter device comparisons. Conclusions These data provide evidence for repeatability and concordance for ranking of glycemic responses, and suggest that factors other than CGM sensors mostly drive within-subject meal categorization. Our findings are critical for identifying sources of variability in glycemic responses for the eventual implementation of precision nutrition. Funding Sources Zoe Global, UK Government Department of Health and Social Care, Wellcome Trust. National Institutes of Health.

Leadership and Motivation as Important Aspects of the International Company’s Corporate Culture

Although the issue of corporate culture has been taken over and addressed in the literature from various perspectives, there are very few researchers about the role of leadership and motivation in it, respectively very few researchers have addressed them as important components of the international company’s corporate culture. The present paper aims to point out that leadership and motivation can be perceived as important aspects of the international company’s corporate culture. The object of the investigation was an international company (situated in Italy) and its five subsidiaries (situated in Italy, Czech Republic, Germany, and Turkey). As the main research method, there was chosen the method of the questionnaire survey, which was attempted by all the company’s employees (totally 270 respondents). The questionnaire was divided into three separate, but logically related parts—leadership, motivation, and corporate culture, and submitted to two groups of respondents—the company’s management and its employees. In total 11 hypotheses were formulated and further evaluated by the methods of Pearson Chi-square Test, Fisher’s Exact Test, Cramer’s V coefficient, Kendall rank correlation coefficient, Eta coefficient, Spearman coefficient, Mann–Whitney U test and Wilcoxon W statistics, Kruskal–Wallis test, and Friedman’s test. The results of the research have proven that leadership and motivation are important parts of the corporate culture.

On the usage of average Hausdorff distance for segmentation performance assessment: hidden error when used for ranking

Average Hausdorff distance is a widely used performance measure to calculate the distance between two point sets. In medical image segmentation, it is used to compare ground truth images with segmentations allowing their ranking. We identified, however, ranking errors of average Hausdorff distance making it less suitable for applications in segmentation performance assessment. To mitigate this error, we present a modified calculation of this performance measure that we have coined “balanced average Hausdorff distance”. To simulate segmentations for ranking, we manually created non-overlapping segmentation errors common in magnetic resonance angiography cerebral vessel segmentation as our use-case. Adding the created errors consecutively and randomly to the ground truth, we created sets of simulated segmentations with increasing number of errors. Each set of simulated segmentations was ranked using both performance measures. We calculated the Kendall rank correlation coefficient between the segmentation ranking and the number of errors in each simulated segmentation. The rankings produced by balanced average Hausdorff distance had a significantly higher median correlation (1.00) than those by average Hausdorff distance (0.89). In 200 total rankings, the former misranked 52 whilst the latter misranked 179 segmentations. Balanced average Hausdorff distance is more suitable for rankings and quality assessment of segmentations than average Hausdorff distance.

Correlations and Hierarchical Clustering investigation between weather and SARS-CoV-2.

Background:: Humanity today faces a global emergency. It is conceivably the greatest crisis of our generation. The coronavirus pandemic, which has many global implications, has led researchers worldwide to seek solutions to this crisis, including the search for effective treatment in the first place. Objective:: This study aims to identify the factors that can have an essential effect on COVID-19 comportment. Having proper management and control of imports of COVID-19 depends on many factors that are highly dependent on a country's sanitary capacity and infrastructure technology. Nevertheless, meteorological parameters can also be a connecting factor to this disease; seines temperature and humidity are compatible with the behavior of a seasonal respiratory virus. Methods:: In this work, we analyze the correlation between weather and the COVID-19 epidemic in Casablanca, the economic capital of Morocco. It is based on the primary analysis of COVID-19 surveillance data from the Ministry of Health of the Kingdom of Morocco and weather data from the meteorological data. Weather factors include minimum temperature (°C), maximum temperature (°C), mean temperature (°C), maximum wind speed (Km/h), humidity (%), and rainfall (mm). The Spearman and Kendall rank correlation test is used for data analysis. Between the weather components. Results:: The mean temperature, maximum temperature (°C) and Humidity were significantly correlated with the COVID-19 pandemic with respectively (r= -0.432, r = -0.480; r=0.402, and p=- 0.212, p= -0.160, and p= &-0.240). Conclusion:: This discovery helps reduce the incidence rate of COVID-19 in Morocco, considering the significant correlation between weather and COVID-19, of about more than 40%.

A Method for Predicting the Remaining Useful Life of Lithium-Ion Batteries Based on Particle Filter Using Kendall Rank Correlation Coefficient

With the wide application of lithium batteries, battery fault prediction and health management have become more and more important. This article proposes a method for predicting the remaining useful life (RUL) of lithium-ion batteries to avoid a series of safety problems caused by continuing to use the battery after reaching its service life threshold. Since the battery capacity is not easy to obtain online, we propose that some measurable parameters should be used in the battery discharge cycle to estimate battery capacity. Then, the estimated capacity is used to replace the measured value of the particle filter (PF) based on the Kendall rank correlation coefficient (KCCPF) to predict the RUL of the lithium batteries. Simulation results show that the proposed method has high prediction accuracy, stability, and practical value.

Imprecision nutrition? Different simultaneous continuous glucose monitors provide discordant meal rankings for incremental postprandial glucose in subjects without diabetes

ABSTRACT Background High postprandial glucose excursions may increase risk for disease. Individuals have widely varying glucose responses to different meals, and precision nutrition approaches often seek to personalize diets to minimize postprandial glycemic responses as measured by continuous glucose monitors (CGMs). However, it is unknown whether different CGM devices result in concordant meal rankings according to postprandial glycemic excursions. Objective We explored whether meal rankings according to postprandial glycemic excursions differ between 2 simultaneously worn CGMs. Methods We collected 27,489 simultaneous measurements from Dexcom G4 Platinum and Abbott Freestyle Libre Pro CGMs during 28 inpatient days in 16 adults without diabetes. Simultaneous glucose measurements obtained for 2 h following 760 ad libitum meals were used to compare within-subject meal rankings between the CGM devices according to their incremental glucose response. Results Postprandial responses to ad libitum meals were highly variable, with the Abbott and Dexcom systems resulting in within-subject incremental mean ± SD glucose CVs of 91.7 ± 1.9% and 94.2 ± 2.7%, respectively. Within-subject meal rankings for incremental glycemic responses were relatively discordant between CGMs, with a mean Kendall rank correlation coefficient of 0.43 ± 0.05. Meals in the bottom compared with those in the top half of incremental glycemic responses ranked by Abbott resulted in 50 ± 10% (P = 0.0002) less glycemic reduction as measured by Dexcom, and vice versa. The missing glycemic reduction by eating meals ranked according to the discordant CGM was inversely correlated with each subject's Kendall rank correlation coefficient (r = −0.95; P < 0.0001). Conclusions Precision nutrition approaches that use CGMs to personalize meal recommendations for minimizing glycemic excursions may be premature given the discordance of within-subject meal rankings between simultaneous CGM devices. More research is needed to clarify the source of this imprecision. This trial was registered at clinicaltrials.gov as NCT03407053.