Dynamic Hybrid Model to Forecast the Spread of COVID-19 Using LSTM and Behavioral Models Under Uncertainty.

To accurately predict the regional spread of coronavirus disease 2019 (COVID-19) infection, this study proposes a novel hybrid model, which combines a long short-term memory (LSTM) artificial recurrent neural network with dynamic behavioral models. Several factors and control strategies affect the virus spread, and the uncertainty arising from confounding variables underlying the spread of the COVID-19 infection is substantial. The proposed model considers the effect of multiple factors to enhance the accuracy in predicting the number of cases and deaths across the top ten most-affected countries at the time of the study. The results show that the proposed model closely replicates the test data, such that not only it provides accurate predictions but it also replicates the daily behavior of the system under uncertainty. The hybrid model outperforms the LSTM model while accounting for data limitation. The parameters of the hybrid models are optimized using a genetic algorithm for each country to improve the prediction power while considering regional properties. Since the proposed model can accurately predict the short-term to medium-term daily spreading of the COVID-19 infection, it is capable of being used for policy assessment, planning, and decision making.

Impact of COVID-19 pandemic on mobility in ten countries and associated perceived risk for all transport modes.

The restrictive measures implemented in response to the COVID-19 pandemic have triggered sudden massive changes to travel behaviors of people all around the world. This study examines the individual mobility patterns for all transport modes (walk, bicycle, motorcycle, car driven alone, car driven in company, bus, subway, tram, train, airplane) before and during the restrictions adopted in ten countries on six continents: Australia, Brazil, China, Ghana, India, Iran, Italy, Norway, South Africa and the United States. This cross-country study also aims at understanding the predictors of protective behaviors related to the transport sector and COVID-19. Findings hinge upon an online survey conducted in May 2020 (N = 9,394). The empirical results quantify tremendous disruptions for both commuting and non-commuting travels, highlighting substantial reductions in the frequency of all types of trips and use of all modes. In terms of potential virus spread, airplanes and buses are perceived to be the riskiest transport modes, while avoidance of public transport is consistently found across the countries. According to the Protection Motivation Theory, the study sheds new light on the fact that two indicators, namely income inequality, expressed as Gini index, and the reported number of deaths due to COVID-19 per 100,000 inhabitants, aggravate respondents' perceptions. This research indicates that socio-economic inequality and morbidity are not only related to actual health risks, as well documented in the relevant literature, but also to the perceived risks. These findings document the global impact of the COVID-19 crisis as well as provide guidance for transportation practitioners in developing future strategies.

Impact of risk attitudes and perception on game theoretic driving interactions and safety.

This study employs game theory to investigate behavioural norms of interaction between drivers at a signalised intersection. The choice framework incorporates drivers' risk perception as well as their risk attitudes. A laboratory experiment is conducted to study the impact of risk attitudes and perception in crossing behaviour at a signalised intersection. The laboratory experiment uses methods from experimental economics to induce incentives and study revealed behaviour. Conflicting drivers are considered to have symmetric disincentives for crashing, to represent a no-fault car insurance environment. The study is novel as it uses experimental data collection methods to investigate perceived risk. Further, it directly integrates perceived risk of crashing with other active drivers into the modelling structure. A theoretical model of intersection crossing behaviour is also developed in this paper. This study shows that right-of-way entitlements assigned without authoritative penalties to at-fault drivers may still improve perceptions of safety. Further, risk aversion amongst drivers attributes to manoeuvring strategies at or below Nash mixed strategy equilibrium. These findings offer a theoretical explanation for interactive manoeuvres that lead to crashes, as opposed to purely statistical methods which provide correlation but not necessarily explanation.

Modelling crash propensity of carshare members.

Carshare systems are considered a promising solution for sustainable development of cities. To promote carsharing it is imperative to make them cost effective, which includes reduction in costs associated to crashes and insurance. To achieve this goal, it is important to characterize carshare users involved in crashes and understand factors that can explain at-fault and not-at fault drivers. This study utilizes data from GoGet carshare users in Sydney, Australia. Based on this study it was found that carshare users who utilize cars less frequently, own one or more cars, have less number of accidents in the past ten years, have chosen a higher insurance excess and have had a license for a longer period of time are less likely to be involved in a crash. However, if a crash occurs, carshare users not needing a car on the weekend, driving less than 1000km in the last year, rarely using a car and having an Australian license increases the likelihood to be at-fault. Since the dataset contained information about all members as well as not-at-fault drivers, it provided a unique opportunity to explore some aspects of quasi-induced exposure. The results indicate systematic differences in the distribution between the not-at-fault drivers and the carshare members based on the kilometres driven last year, main mode of travel, car ownership status and how often the car is needed. Finally, based on this study it is recommended that creating an incentive structure based on training and experience (based on kilometres driven), possibly tagged to the insurance excess could improve safety, and reduce costs associated to crashes for carshare systems.