Novel navigation applications provide a driving behavior score for each finished trip to promote safe driving, which is mainly based on experts’ domain knowledge. In this paper, with automobile insurance claims data and associated telematics car driving data, we propose a supervised driving risk scoring neural network model. This one-dimensional convolutional neural network takes time series of individual car driving trips as input and returns a risk score in the unit range of (0,1). By incorporating credibility average risk score of each driver, the classical Poisson generalized linear model for automobile insurance claims frequency prediction can be improved significantly. Hence, compared with non-telematics-based insurers, telematics-based insurers can discover more heterogeneity in their portfolio and attract safer drivers with premiums discounts.

In the auto insurance industry, a Bonus-Malus System (BMS) is commonly used as a posteriori risk classification mechanism to set the premium for the next contract period based on a policyholder's claim history. Even though the recent literature reports evidence of a significant dependence between frequency and severity, the current BMS practice is to use a frequency-based transition rule while ignoring severity information. Although Oh et al. [(2020). Bonus-Malus premiums under the dependent frequency-severity modeling. Scandinavian Actuarial Journal 2020(3): 172–195] claimed that the frequency-driven BMS transition rule can accommodate the dependence between frequency and severity, their proposal is only a partial solution, as the transition rule still completely ignores the claim severity and is unable to penalize large claims. In this study, we propose to use the BMS with a transition rule based on both frequency and size of claim, based on the bivariate random effect model, which conveniently allows dependence between frequency and severity. We analytically derive the optimal relativities under the proposed BMS framework and show that the proposed BMS outperforms the existing frequency-driven BMS. Later, numerical experiments are also provided using both hypothetical and actual datasets in order to assess the effect of various dependencies on the BMS risk classification and confirm our theoretical findings.

The green peafowl Pavo muticus is a highly threatened galliform species that was historically distributed widely across South-east Asia. Evidence shows a recent population decline and range contraction for this species, linked with habitat degradation and over-exploitation. This study aimed to determine the current known distribution across mainland South-east Asia and investigate potential habitat that could host remaining viable populations and contribute to the long-term survival of the species. We used locations from historical and recent records and habitat variables from a geographical information system database to model the probability of occurrence and classify key localities according to their relative importance for the species. Our results showed that the green peafowl probably occurs in less than 16% of its historical range across mainland South-east Asia and that remaining locations are fragmented. Four confirmed and two potential stronghold populations were identified for the species, based on the localities with high capacity to contribute to its long-term survival in large contiguous patches. These were in central Myanmar, western and northern Thailand, eastern Cambodia/south-central Viet Nam and northern Cambodia/southern Lao. Threats vary amongst countries, with continued habitat loss and degradation in many areas and hunting particularly acute in Viet Nam, Cambodia and Lao. Most of the remaining populations are in protected areas but the protection level varies widely. We propose conservation actions for each stronghold population, in accordance with the nature of the threats and protection level in each area, to prevent the local extinction of this species.

We consider compound Poisson claims reserving models applied to the paid claims and to the number of payments run-off triangles. We extend the standard Poisson-gamma assumption to account for over-dispersion in the payment counts and to account for various mean and variance structures in the individual payments. Two generalized linear models are applied consecutively to predict the unpaid claims. A bootstrap is used to estimate the mean squared error of prediction and to simulate the predictive distribution of the unpaid claims. We show that the extended compound Poisson models make reasonable predictions of the unpaid claims.

To manage or control nonindigenous species (NIS), we need to know where they are located in the landscape. However, many natural areas are large, making it unfeasible to inventory the entire area and necessitating surveys to be performed on smaller areas. Provided appropriate survey methods are used, probability of occurrence predictions and maps can be generated for the species and area of interest. The probability maps can then be used to direct further sampling for new populations or patches and to select populations to monitor for the degree of invasiveness and effect of management. NIS occurrence (presence or absence) data were collected during 2001 to 2003 using transects stratified by proximity to rights-of-way in the northern range of Yellowstone National Park. In this study, we evaluate the use of environmental and remotely sensed (LANDSAT Enhanced Thematic Mapper +) data, separately and combined, for developing probability maps of three target NIS occurrence. Canada thistle, dalmation toadflax, and timothy were chosen for this study because of their different dispersal mechanisms and frequencies, 5, 3, and 23%, respectively, in the surveyed area. Data were analyzed using generalized linear regression with logit link, and the best models were selected using Akaike's Information Criterion. Probability of occurrence maps were generated for each target species, and the accuracies of the predictions were assessed with validation data excluded from the model fitting. Frequencies of occurrence of the validation data were calculated and compared with predicted probabilities. Agreement between the observed and predicted probabilities was reasonably accurate and consistent for timothy and dalmation toadflax but less so for Canada thistle.

Species richness of macrophytes of artificial and natural water bodies covering rivers, streams, canals and habitats with standing water was studied in two Central European biogeographical regions, Pannonian and Carpathian, represented by two model areas (the Borská nížina Lowland and the Turčianska Kotlina Basin). We found that: (i) the total number of macrophytes was higher in artificial water bodies compared to natural aquatic habitats in both regions and differences were statistically significant (P<0.05); and (ii) species richness of macrophytes is relatively low in both regions; slightly higher mean number was found in the Pannonian region (3.53) compared with the Carpathian region (3.06). Effects of environmental characteristics on species richness studied by generalized linear model (GLM) showed that explained variances of GLM were similar in both regions. The main drivers of the diversity pattern were those connected with the hydrology of water bodies, such as substrate characteristics and turbidity, less frequent were chemical characteristics, such as water conductivity and N-contents, whereas landscape characteristics manifested the smallest impact.

The ichthyoplankton of Galway Bay was sampled over the spring and summer of 2007 and 2008 to determine what environmental factors influenced the distribution of larval and early juvenile fish. A total of 549 fish representing 27 taxa were caught. Catches decreased throughout each sampling season, and were generally poor in 2007. Catches were numerically dominated by sprat (Sprattus sprattus), dab (Limanda limanda), sand eel (Hyperoplus immaculatus) and whiting (Merlangius merlangus). Environmental factors driving distribution of fish were modelled using a binomial generalized linear model. The strength and direction of wind in the five days preceding sampling was the only significant environmental factor. Sustained onshore winds increased the probability of encountering larval fish in the areas sampled. The rainfall, tidal state, tow depth and wind conditions during sampling were not significant. Among the parameters measured it appears that onshore winds sufficiently strong to overcome the residual circulation are the main physical driver for the distribution of larval fish in Galway Bay.

Research in epidemiology may be concerned with assessing risk factors for complex health issues described by several variables. Moreover, epidemiological data are usually organized in several blocks of variables, consisting of a block of variables to be explained and a large number of explanatory variables organized in meaningful blocks. Usual statistical procedures such as generalized linear models do not allow the explanation of a multivariate outcome, such as a complex disease described by several variables, with a single model. Moreover, it is not easy to take account of the organization of explanatory variables into blocks. Here we propose an innovative method in the multiblock modelling framework, called multiblock redundancy analysis, which is designed to handle most specificities of complex epidemiological data. Overall indices and graphical displays associated with different interpretation levels are proposed. The interest and relevance of multiblock redundancy analysis is illustrated using a dataset pertaining to veterinary epidemiology.

As a consequence of pointing out an ambiguity in Renshaw (1994), we show that the Overdispersed Poisson model cannot be generated by random independent intensities. Hence Pearson's chi-square-based estimate is normally unsuitable for GLM (Generalized Linear Model) log link claim frequency analysis in insurance. We propose a new dispersion parameter estimate in the GLM Tweedie model for risk premium. This is better than the Pearson estimate, if there are sufficiently many claims in each tariff cell. Simulation results are given showing the differences between it and the Pearson estimate.

Among the most endangered tropical ecosystems, tropical dry forests are threatened by degradation that includes edge effects arising from perturbations such as the creation and maintenance of roads and other clearings. While much is known about these adverse effects on tree communities in tropical moist forests, similar effects in tropical dry forests are little understood. This paper examines the relationship between roads, road-related exotic plant invasion and tree community change in a tropical dry forest in southern India. Forty pairs of roadside and interior plots across four factorial combinations of road width (wide and narrow) and understorey type (native and exotic) were sampled. Tree death and extant tree community composition were compared using generalized linear models and similarity analyses. Tree death near roads was more than double that away from them, suggesting that roads may increase tree death in these forests. The interactive effect of understorey type (exotic or native) and road width on tree death was significant, with highest tree death near wide roads bordered by exotic understorey. Conversely, tree community composition was influenced by road width and understorey type, but not by proximity to roads. Creation and maintenance of roads for forest management may have serious implications for tree communities in tropical dry forests and should thus be minimized. Exotic plants may also be important contributors to increasing tree death, and further research on their impacts, particularly into underlying mechanisms, is critical to the long-term conservation of tropical dry forest communities.

A large database on the transmission of a stabilate of the Theileria parva Muguga stock from one breed of cattle using two stocks of Rhipicephalus appendiculatus, Muguga and Ol Pejeta was developed and analysed. Factors associated with the ticks and cattle, and the infections developing in cattle were studied in relation to the infection variables in the tick batches harvested daily from cattle. Generalized Linear Interactive Modelling (GLIM) was used to determine the importance of factors and interactions in influencing the levels of tick infection variables using Type I and Type III sums of squares analyses. Analysis of the 6 variables, prevalence (percentage of ticks infected), abundance (mean number of infected salivary gland acini per tick examined) and intensity (mean number of infected salivary gland acini per infected tick) in batches of 30 male and 30 female ticks showed that 24 covariates, factors or interactions had a significant effect (P> 0·05). Certain covariates and factors were particularly important for all 6 tick infection variables; parasitaemia of animal on the day of tick harvest, stabilate dilution administered to animal, month in which tick batch was harvested, minimum packed cell volume of animal over the sampling period, age of animal, and the minimum leukocyte count of the animal over the sampling period. The GLIM analyses were found to be a useful tool in identifying factors that influence infection levels and in devising methods of producing tick batches with more predictable infections.

The humpback whale (Megaptera novaeangliae) population that uses Abrolhos Bank, off the east coast of Brazil as a breeding ground is increasing. To describe temporal changes in the relative abundance of humpback whales around Abrolhos, seven years (1998–2004) of whale count data were collected during July through to November. During one-hour-scans, observers determined group size within 9.3 km (5 n.m.) of a land-based observing station. A total of 930 scans, comprising 7996 sightings of adults and 2044 calves were analysed using generalized linear models that included variables for time of day, day of the season, years and two-way interactions as possible predictors. The pattern observed was the gradual build-up and decline in whale counts within seasons. Patterns and peaks of adult and calf counts varied among years. Although fluctuation was observed, there was generally an increasing trend in adult counts among years. Calf counts increased only in 2004. These fluctuations may have been caused by some environmental conditions in humpback whales' summering grounds and also by changes in spatial–temporal concentrations in Abrolhos Bank. The general pattern observed within the study area mirrored what was observed in the whole Abrolhos Bank. Knowledge of the consistency with which humpback whales use this important nursing area should prove beneficial for designing future monitoring programmes especially related to whale watching activities around Abrolhos Archipelago.