The aim of the paper is to derive a simple, implementable machine learning method for general insurance losses. An algorithm for learning a general insurance loss triangle is developed and justified. An argument is made for applying support vector regression (SVR) to this learning task (in order to facilitate transparency of the learning method as compared to more “black-box” methods such as deep neural networks), and SVR methodology derived is specifically applied to this learning task. A further argument for preserving the statistical features of the loss data in the SVR machine is made. A bespoke kernel function that preserves the statistical features of the loss data is derived from first principles and called the exponential dispersion family (EDF) kernel. Features of the EDF kernel are explored, and the kernel is applied to an insurance loss estimation exercise for homogeneous risk of three different insurers. Results of the cumulative losses and ultimate losses predicted by the EDF kernel are compared to losses predicted by the radial basis function kernel and the chain-ladder method. A backtest of the developed method is performed. A discussion of the results and their implications follows.

Understanding the effects of predicted rising sea levels, combined with changes in precipitation and freshwater inflow on key estuarine ecosystem engineers such as the eastern oyster would provide critical information to inform restoration design and predictive models. Using oyster ladders with shell bags placed at three heights to capture a range of inundation levels, oyster growth of naturally recruited spat was monitored over the course of 6 months. Oyster numbers and shell heights were consistently highest in bottom and mid bags experiencing greater than 50% inundation (mid: 63 ± 7%; bottom: 95 ± 3%). Identifying thresholds for optimal oyster growth and survival to enhance restoration engineering would require finer scale evaluation of inundation levels.

The President (Dr J. Taylor, F.F.A.): Good evening from Edinburgh. Today marks the first-ever presidential address of the merged Institute and Faculty of Actuaries (IFoA) to take place outside of London.

During the last months and following the implementation of containment measures in the context of coronavirus disease 2019 (COVID-19) pandemic, the number of new human immunodeficiency virus (HIV) diagnoses radically decreased in Liege AIDS Reference Center, Belgium. The number of HIV screening tests has also dramatically dropped down to an unprecedented level. This decline of HIV diagnosis is caused by missed diagnoses of individuals infected before the establishment of such measures and to the reduction of high-risk sexual behaviours during the COVID-19 pandemic.

Bordetella bronchiseptica is a potential zoonotic pathogen, which mainly causes respiratory diseases in humans and a variety of animal species. B. bronchiseptica is one of the important pathogens isolated from rabbits in Fujian Province. However, the knowledge of the epidemiology and characteristics of the B. bronchiseptica in rabbits in Fujian Province is largely unknown. In this study, 219 B. bronchiseptica isolates recovered from lung samples of dead rabbits with respiratory diseases in Fujian Province were characterised by multi-locus sequencing typing, screening virulence genes and testing antimicrobial susceptibility. The results showed that the 219 isolates were typed into 11 sequence types (STs) including five known STs (ST6, ST10, ST12, ST14 and ST33) and six new STs (ST88, ST89, ST90, ST91, ST92 and ST93) and the ST33 (30.14%, 66/219), ST14 (26.94%, 59/219) and ST12 (16.44%, 36/219) were the three most prevalent STs. Surprisingly, all the 219 isolates carried the five virulence genes (fhaB, prn, cyaA, dnt and bteA) in the polymerase chain reaction screening. Moreover, the isolates were resistant to cefixime, ceftizoxime, cefatriaxone and ampicillin at rates of 33.33%, 31.05%, 11.87% and 3.20%, respectively. This study showed the genetic diversity of B. bronchiseptica in rabbits in Fujian Province, and the colonisation of the human-associated ST12 strain in rabbits in Fujian Province. The results might be useful for monitoring the epidemic strains, developing preventive methods and preventing the transmission of epidemic strains from rabbits to humans.

In this paper, we mainly study a class of small deviation theorems for Markov chains indexed by an infinite tree with uniformly bounded degree in Markovian environment. Firstly, we give the definition of Markov chains indexed by a tree with uniformly bounded degree in random environment. Then, we introduce the some lemmas which are the basis of the results. Finally, a class of small deviation theorems for functionals of random fields on a tree with uniformly bounded degree in Markovian environment is established.

There is growing interest in quantifying attitudes towards autistic people, however there is relatively little research on psychometric properties of the only existing measure and its ability to predict engagement with people with autism. To begin addressing these issues, we compared three scales measuring attitudes towards autistic people following the development of two new measures. Exploratory factor analysis, across two datasets, revealed that the factor-structure of an established 16-item scale is unclear. Further, its predictive validity of intended engagement with autistic people was comparable to our novel and psychometrically robust 1- and 4-item measures of attitudes towards autistic people. We therefore conclude that a 1- or 4-item scale is sufficient to measure general attitudes towards autistic people in future research. Equally, we propose that additional research is required to develop measures that are grounded in theoretical models of attitude formation and therefore distinguish between different components of attitudes.

The ongoing coronavirus disease 2019 (COVID-19) pandemic is of global concern and has recently emerged in the US. In this paper, we construct a stochastic variant of the SEIR model to estimate a quasi-worst-case scenario prediction of the COVID-19 outbreak in the US West and East Coast population regions by considering the different phases of response implemented by the US as well as transmission dynamics of COVID-19 in countries that were most affected. The model is then fitted to current data and implemented using Runge-Kutta methods. Our computation results predict that the number of new cases would peak around mid-April 2020 and begin to abate by July provided that appropriate COVID-19 measures are promptly implemented and followed, and that the number of cases of COVID-19 might be significantly mitigated by having greater numbers of functional testing kits available for screening. The model is also sensitive to assigned parameter values and reflects the importance of healthcare preparedness during pandemics.

Purpose: The novel coronavirus (severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2)) first appeared in Wuhan, China, in December 2019, and rapidly spread across the globe. Since most respiratory viruses are known to show a seasonal pattern of infection, it has been hypothesised that SARS-CoV-2 may be seasonally dependent as well. The present study looks at a possible effect of atmospheric temperature, which is one of the suspected factors influencing seasonality, on the evolution of the pandemic. Basic procedures: Since confirming a seasonal pattern would take several more months of observation, we conducted an innovative day-to-day micro-correlation analysis of nine outbreak locations, across four continents and both hemispheres, in order to examine a possible relationship between atmospheric temperature (used as a proxy for seasonality) and outbreak progression. Main findings: There was a negative correlation between atmospheric temperature variations and daily new cases growth rates, in all nine outbreaks, with a median lag of 10 days. Principal conclusions: The results presented here suggest that high temperatures might dampen SARS-CoV-2 propagation, while lower temperatures might increase its transmission. Our hypothesis is that this could support a potential effect of atmospheric temperature on coronavirus disease progression, and potentially a seasonal pattern for this virus, with a peak in the cold season and rarer occurrences in the summer. This could guide government policy in both the Northern and Southern hemispheres for the months to come.

Given extensive research underscoring the deleterious effects of bullying on youth adjustment, anti-bullying policies and programming are critical public health priorities. However, strategies that increase public support for anti-bullying causes are not well understood. This experiment assessed the influence of “bullying messaging” on support for anti-bullying policies. Specifically, I investigated whether learning about the health consequences of bullying, as opposed to its prevalence or educational impact, increased individuals’ support of anti-bullying policies. Participants (n = 329) were randomly assigned to one of four conditions where they read a brief summary about bullying research; conditions varied by whether the research documented the: a) prevalence of bullying b) mental health consequences of bullying c) physical health consequences of bullying or d) academic consequences of bullying. Results indicated that participants endorsed high levels of support for anti-bullying policies, regardless of experimental condition, and that policies aimed at increasing K-12 mental health resources were most supported.

Acute haemorrhagic conjunctivitis is a highly contagious eye disease, the prediction of acute haemorrhagic conjunctivitis is very important to prevent and grasp its development trend. We use the exponential smoothing model and the seasonal autoregressive integrated moving average (SARIMA) model to analyse and predict. The monthly incidence data from 2004 to 2017 were used to fit two models, the actual incidence of acute haemorrhagic conjunctivitis in 2018 was used to validate the model. Finally, the prediction effect of exponential smoothing is best, the mean square error and the mean absolute percentage error were 0.0152 and 0.1871, respectively. In addition, the incidence of acute haemorrhagic conjunctivitis in Chongqing had a seasonal trend characteristic, with the peak period from June to September each year.

Different countries have adopted strategies for the early detection of SARS-CoV-2 since the declaration of community transmission by the World Health Organization (WHO) and timely diagnosis has been considered one of the major obstacles for surveillance and healthcare. Here, we report the increase of the number of laboratories to COVID-19 diagnosis in Brazil. Our results demonstrate an increase and decentralisation of certified laboratories, which does not match the much higher increase in the number of COVID-19 cases. Also, it becomes clear that laboratories are irregularly distributed over the country, with a concentration in the most developed state, São Paulo.

We establish a fundamental property of bivariate Pareto records for independent observations uniformly distributed in the unit square. We prove that the asymptotic conditional distribution of the number of records broken by an observation given that the observation sets a record is Geometric with parameter 1/2.

To provide comprehensive information on the epidemiology and burden of respiratory syncytial virus hospitalisation (RSVH) in preterm infants, a pooled analysis was undertaken of seven multicentre, prospective, observational studies from across the Northern Hemisphere (2000–2014). Data from all 320–356 weeks' gestational age (wGA) infants without comorbidity were analysed. RSVH occurred in 534/14 504 (3.7%) infants; equating to a rate of 5.65 per 100 patient-seasons, with the rate in individual wGA groups dependent upon exposure time (P = 0.032). Most RSVHs (60.1%) occurred in December–January. Median age at RSVH was 88 days (interquartile range (IQR): 54–159). Respiratory support was required by 82.0% of infants: oxygen in 70.4% (median 4 (IQR: 2–6) days); non-invasive ventilation in 19.3% (median 3 (IQR: 2–5) days); and mechanical ventilation in 10.2% (median 5 (IQR: 3–7) days). Intensive care unit admission was required by 17.9% of infants (median 6 days (IQR: 2–8) days). Median overall hospital length of stay (LOS) was 5 (IQR: 3–8) days. Hospital resource use was similar across wGA groups except for overall LOS, which was shortest in those born 35 wGA (median 3 vs. 4–6 days for 32–34 wGA; P < 0.001). Strategies to reduce the burden of RSVH in otherwise healthy 32–35 wGA infants are indicated.

Cyclopentadithiophene (CPDT), a Csp3-bridged bithiophene heteroaromatic unit, displays interesting properties when it is embedded in the repeating units of π-conjugated polymers, and they are applied in organic electronics devices. Common synthetic routes to CPDT-derived polymers rely on toxic methodologies whilst alternative non-toxic strategies such as the Suzuki-Miyaura reaction have been less studied. In this report we demonstrate that the use of a N-methyliminodiacetic acid (MIDA) boronate ester-derived CPDT monomer allows the efficient formation of poly(cyclopentadithiophene) homopolymer under Suzuki-Miyaura cross-coupling reaction conditions. Thus, the use of MIDA boronate esters might be extended to other organic units to design and construct a plethora of π-conjugated polymers.

Financial products are priced using risk-neutral expectations justified by hedging portfolios that (as accurate as possible) match the product’s payoff. In insurance, premium calculations are based on a real-world best-estimate value plus a risk premium. The insurance risk premium is typically reduced by pooling of (in the best case) independent contracts. As hybrid life insurance contracts depend on both financial and insurance risks, their valuation requires a hybrid valuation principle that combines the two concepts of financial and actuarial valuation. The aim of this paper is to present a novel three-step projection algorithm to valuate hybrid contracts by decomposing their payoff in three parts: a financial, hedgeable part, a diversifiable actuarial part, and a residual part that is neither hedgeable nor diversifiable. The first two parts of the resulting premium are directly linked to their corresponding hedging and diversification strategies, respectively. The method allows for a separate treatment of unsystematic, diversifiable mortality risk and systematic, aggregate mortality risk related to, for example, epidemics or population-wide improvements in life expectancy. We illustrate our method in the case of CAT bonds and a pure endowment insurance contract with profit and compare the three-step method to alternative valuation operators suggested in the literature.

During the past decade, genetics research has allowed scientists and clinicians to explore the human genome in detail and reveal many thousands of common genetic variants associated with disease. Genetic risk scores, known as polygenic risk scores (PRSs), aggregate risk information from the most important genetic variants into a single score that describes an individual’s genetic predisposition to a given disease. This article reviews recent developments in the predictive utility of PRSs in relation to a person’s susceptibility to breast cancer and coronary artery disease. Prognostic models for these disorders are built using data from the UK Biobank, controlling for typical clinical and underwriting risk factors. Furthermore, we explore the possibility of adverse selection where genetic information about multifactorial disorders is available for insurance purchasers but not for underwriters. We demonstrate that prediction of multifactorial diseases, using PRSs, provides population risk information additional to that captured by normal underwriting risk factors. This research using the UK Biobank is in the public interest as it contributes to our understanding of predicting risk of disease in the population. Further research is imperative to understand how PRSs could cause adverse selection if consumers use this information to alter their insurance purchasing behaviour.

Erdős, Gyárfás and Pyber showed that every r-edge-coloured complete graph Kn can be covered by 25 r2 log r vertex-disjoint monochromatic cycles (independent of n). Here we extend their result to the setting of binomial random graphs. That is, we show that if $p= p(n) = \Omega(n^{-1/(2r)})$, then with high probability any r-edge-coloured G(n, p) can be covered by at most 1000r4 log r vertex-disjoint monochromatic cycles. This answers a question of Korándi, Mousset, Nenadov, Škorić and Sudakov.

The President (Mr J. Constantinou, F.F.A.): Hello and welcome to Staple Inn, the spiritual home of the Institute and Faculty of Actuaries (IFoA). I feel privileged to be standing before you as the President of the IFoA.