論文

Scientists often collect samples on characteristics of different observation units and wonder whether those characteristics have similar distributional structure. We consider methods to find homogeneous subpopulations in a multidimensional space using regression tree and clustering methods for distributions of a population characteristic. We present a new methodology to estimate a standardized measure of distance between clusters of distributions and for hierarchical testing to find the minimal homogeneous or near-homogeneous tree structure. In addition, we introduce hierarchical clustering with adjacency constraints, which is useful for clustering georeferenced distributions. We conduct simulation studies to compare clustering performance with three measures: Modified Jensen–Shannon divergence (MJS), Earth Mover’s distance and Cramér–von Mises distance to validate the proposed testing procedure for homogeneity. As a motivational example, we introduce georeferenced yellowfin tuna fork length data collected from the catch of purse-seine vessels that operated in the eastern Pacific Ocean. Hierarchical clustering, with and without spatial adjacency constraints, and regression tree methods were applied to the density estimates of length. While the results from the two methods showed some similarities, hierarchical clustering with spatial adjacency produced a more flexible partition structure, without requiring additional covariate information. Clustering with MJS produced more stable results than clustering with the other measures.

For large-scale tropical tuna purse-seine fisheries, it is prohibitively costly to obtain adequate sampling coverage to estimate fleet-level catch composition solely from sample data. Logbook or observer data, with complete fleet coverage, are often available but may be considered unreliable for species composition. Previous studies have developed models, trained with sample data, to predict set-level species compositions based on environmental and operational covariates. Here, models were developed to predict well-level species composition from uncorrected observer data and covariates affecting the observers’ view of the catch during loading, with port-sampling data as the response variable. The analysis used paired, well-level data from sets made on floating objects by the Eastern Pacific Ocean tuna purse-seine fleet during 2023–2024. Results indicated that, overall, observer data proportions of bigeye (BET) and yellowfin tunas tended to be greater than the model-estimated proportions, with the opposite occurring for skipjack tuna (SKJ). However, vessel effects sometimes modified these tendencies. Model complexity was greatest for BET and least for SKJ. For BET, observer data proportions and model-estimated proportions were more similar when the vessel had a hopper. They were also more similar in 2023 as compared to 2024, suggesting sample data for bias adjustments should be collected annually. The approach shows potential for predicting the species composition of unsampled wells.

Aims: Long-term exposure to fine particulate matter (PM2.5) is causally associated with mortality and cardiovascular disease. However, in terms of cardiovascular cause-specific outcomes, there are fewer studies about stroke than about coronary heart disease, particularly in Asia. Furthermore, there remains uncertainty regarding the PM2.5-respiratory disease association. We examined whether long-term exposure to PM2.5 is associated with all-cause, cardiovascular and respiratory disease mortality in Japan. Methods: We used data of 46,974 participants (19,707 men; 27,267 women), who were enrolled in 2009 and followed up until 2019, in a community-based prospective cohort study (the second cohort of the Ibaraki Prefectural Health Study). We estimated PM2.5 concentrations using the inverse distance weighing methods based on ambient air monitoring data, and assigned each participant to administrative area level concentrations. A Cox proportional hazard model was applied to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of mortality. Results: During the average follow-up of 10 years, we confirmed 2,789 all-cause deaths. All outcomes including stroke mortality did not significantly increase as the PM2.5 concentration increased. For non-malignant respiratory disease mortality, the multivariable adjusted HR per 1 µg/m3 increase in the PM2.5 concentration was 1.09 (95% CI = 0.97–1.23). Conclusions: In this population exposed to PM2.5 at concentrations of 8.3–13.1 µg/m3, there was no evidence that long-term exposure to PM2.5 had adverse effects on mortality. Weak evidence of positive association observed for non-malignant respiratory disease mortality needs further studies in other populations.

The peripubertal growth pattern of lung function remains underexplored in relation to height growth. This study aimed to first clarify the relationship between the age at peak growth velocity in lung function variables and the age at peak height velocity (APHV) and second identify sex differences in lung function growth patterns. Lung function and height were measured annually in children aged 9–15 years (elementary schools, N = 1307; junior high schools, N = 792) from 2011 to 2018. Children were categorized quarterly according to APHV, using the Super Imposition by Translation and Rotation model. The age at peak growth velocity for forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1) lagged behind APHV by 2–12 months. The later the APHV, the greater the numerical lag, although this was not significant. In males, but not females, the trajectory of FEV1/FVC values gradually decreased to reach the lowest levels and then gradually increased with age (U-shaped curve) in all quartiles. Both FVC and FEV1 overwhelmed in males compared with those in females when the height exceeded 150–160 cm. Our results highlight significant variability in peripubertal lung growth with height and sex-related differences in the growth of airways and parenchymal components.

Rationale: Epidemiological evidence indicates that ambient exposure to particulate matter <2.5 μm in aerodynamic diameter (PM2.5) has adverse effects on lung function growth in children, but it is not actually clear whether exposure to low-level PM2.5 results in long-term decrements in lung function growth in preto early-adolescent schoolchildren. Objectives: To examine long-term effects of PM2.5 within the 4-year average concentration range of 10-19 μg/m3 on lung function growth with repeated measurements of lung function tests. Methods: Longitudinal analysis of 6,233 lung function measurements in 1,466 participants aged 8-12 years from 16 school communities in 10 cities around Japan, covering a broad area of the country to represent concentration ranges of PM2.5, was done with a multilevel linear regression model. Forced expiratory volume in 1 second, forced vital capacity (FVC), and maximal expiratory flow at 50% of FVC were used as lung function indicators to examine the effects of 10-μg/m3 increases in the PM2.5 concentration on relative growth per each 10-cm increase in height. Results: The overall annual mean PM2.5 level was 13.5 μg/m3 (range, 10.4-19.0 μg/m3). We found no association between any of the lung function growth indicators and increases in PM2.5 levels in children of either sex, even after controlling for potential confounders. Analysis with two-pollutant models with O3 or NO2 did not change the null results. Conclusions: This nationwide longitudinal study suggests that concurrent, long-term exposure to PM2.5 at concentrations ranging from 10.4 to 19.0 μg/m3 has little effect on lung function growth in preadolescent boys or pre- to early-adolescent girls.

Abstract Missing data can introduce biases in the estimation of the indicator of interest if appropriate adjustments are not made. The case of Missing Not at Random (MNAR), a missing mechanism in which the missingness also depends on the missing values themselves, has been under-explored. When an outcome has MNAR data, one method to estimate the population mean of the outcome is using the extended propensity score. This method first estimates the extended propensity score, which is the missing probability conditional on the outcome and covariates. Then, the population mean of the outcome is estimated using these estimates. In this paper, we discuss which variables should be included in or excluded from the extended propensity score model to obtain an unbiased estimate of the population mean with small standard errors. First, we show which covariates, at a minimum, should be included in the model of missing probability so that the population mean estimator of the outcome is consistent. Next, we show that the inclusion of some covariates in the missing probability model results in a large variance of the population mean estimates even if they explain the missing probability well. Then, we verify these arguments using simulation experiments and argue that to obtain unbiased, small-variance estimates of the population mean, it is desirable to include only those covariates necessary for consistency. This study allows us to obtain such estimates when the outcome is MNAR and adjusted by the extended propensity score.

When we study treatment effects in observational studies, we often face two problems: confounding and missing data. To properly estimate the treatment effects with missing data, we need to solve both problems simultaneously. When the outcome is missing not at random (MNAR), existing methods for estimating treatment effects require outcome modeling. However, identifying an appropriate model for a variable with MNAR is difficult. In this paper, we propose a method to estimate the average treatment effect (ATE), the average treatment effect on the treated (ATT), and the average treatment effect on the untreated (ATU) using an instrumental variable for missing data without modeling the outcome when it is MNAR. Here, the instrumental variable is independent of the potential outcome and associated with the missingness, when conditioned on the covariates. We show that the proposed estimators for ATE, ATT, and ATU are consistent and asymptotically normal. In addition, we demonstrate that the proposed estimators perform better in terms of bias and root-mean-squared error than other commonly used missing data adjustment methods. Moreover, the variance formula of the proposed estimator is verified through simulated experiments. This study allows us to estimate causal effects without bias and without modeling the outcome when the outcome is MNAR.

Nocturnal fireflies sometimes use intricate bioluminescent signal systems for sexual communication. In this study, we examined flash signals and mate recognition in the Asian firefly, Aquatica lateralis, under natural field conditions. We found that the flash pattern of females changes after copulation, from simple short flashes to flashes with longer duration and flickering. To understand the functions of flickering, we video-recorded and analyzed the flashes of sedentary males, receptive females, and mated females. The results showed that the flashes of these three adult phases can be discriminated from each other by two parameters, flash duration and flicker intensity, with little overlap. Male attraction experiments using an artificial LED device termed ‘e-firefly’ confirmed that flying and sedentary males are attracted to flashes with shorter durations and lower flicker intensities. The range of attraction success was much wider for flying males and narrower for sedentary males, and the latter was close to the range of receptive female’s flashes. These findings suggest that in addition to flash duration, flicker intensity is a flash signal parameter of mate recognition in A. lateralis males.

© 2018 John Wiley & Sons Ltd Recent large fluctuations in an index of relative abundance for the silky shark in the eastern Pacific Ocean have called into question its reliability as a population indicator for management. To investigate whether these fluctuations were driven by environmental forcing rather than true changes in abundance, a Pacific-wide approach was taken. Data collected by observers aboard purse-seine vessels fishing in the equatorial Pacific were used to compute standardized trends in relative abundance by region, and where possible, by shark size category as a proxy for life stage. These indices were compared to the Pacific Decadal Oscillation (PDO), an index of Pacific Ocean climate variability. Correlation between silky indices and the PDO was found to differ by region and size category. The highest correlations by shark size category were for small (<90 cm total length [TL]) and medium (90–150 cm TL) sharks from the western region of the equatorial eastern Pacific (EP) and from the equatorial western Pacific. This correlation disappeared in the inshore EP. Throughout, correlations with the PDO were generally lower for large silky sharks (>150 cm TL). These results are suggestive of changes in the small and medium silky indices being driven by movement of juvenile silky sharks across the Pacific as the eastern edge of the Indo-Pacific Warm Pool shifts location with ENSO events. Lower correlation of the PDO with large shark indices may indicate that those indices were less influenced by environmental forcing and therefore potentially less biased with respect to monitoring population trends.

© 2019, © 2019 Informa UK Limited, trading as Taylor & Francis Group. Determining risk contributions of unit exposures to portfolio-wide economic capital is an important task in financial risk management. Computing risk contributions involves difficulties caused by rare-event simulations. In this study, we address the problem of estimating risk contributions when the total risk is measured by value-at-risk (VaR). Our proposed estimator of VaR contributions is based on the Metropolis-Hasting (MH) algorithm, which is one of the most prevalent Markov chain Monte Carlo (MCMC) methods. Unlike existing estimators, our MH-based estimator consists of samples from the conditional loss distribution given a rare event of interest. This feature enhances sample efficiency compared with the crude Monte Carlo method. Moreover, our method has consistency and asymptotic normality, and is widely applicable to various risk models having a joint loss density. Our numerical experiments based on simulation and real-world data demonstrate that in various risk models, even those having high-dimensional (≈500) inhomogeneous margins, our MH estimator has smaller bias and mean squared error when compared with existing estimators.

© 2020, Springer Science+Business Media, LLC, part of Springer Nature. Management of large-scale pelagic fisheries relies heavily on fishery data to provide information on tuna population status because, for widely distributed populations, the cost of collecting survey data is often prohibitively high. However, fishery data typically do not provide direct information on interactions among fishing vessels, and thus methods of analysis often assume that vessels operate independently, despite the belief that cooperative fishing occurs. Cluster analysis methods were applied to daily vessel location data collected by onboard fisheries observers to identify groups of tuna purse-seine vessels searching for fish close to each other in space. Some vessel groups were found to reoccur through time, both on daily and monthly or longer time scales. This temporal persistence and reoccurrence are interpreted as an indication of cooperative fishing. Results indicate that there may be multiple layers of vessel interactions, from groups of a few vessels to networks of larger numbers of vessels. The use of reoccurring vessel group characteristics to study the temporal and spatial persistence of areas of high tuna abundance is discussed.