ABSTRACT
In recent years, bikesharing systems have become increasingly popular as affordable and sustainable micromobility solutions. Advanced mathematical models such as machine learning are required to generate good forecasts for bikeshare demand. To this end, this study proposes a machine learning modeling framework to estimate hourly demand in a large-scale bikesharing system. Two Extreme Gradient Boosting models were developed: one using data from before the COVID-19 pandemic (March 2019 to February 2020) and the other using data from during the pandemic (March 2020 to February 2021). Furthermore, a model interpretation framework based on SHapley Additive exPlanations was implemented. Based on the relative importance of the explanatory variables considered in this study, share of female users and hour of day were the two most important explanatory variables in both models. However, the month variable had higher importance in the pandemic model than in the pre-pandemic model. © 2023 Informa UK Limited, trading as Taylor & Francis Group.
ABSTRACT
In this paper, we conduct an empirical study of the impact of language distance on Thailand's foreign trade on the basis of trade data from Thailand and its 21 trade partners within 17 years, by applying the trade gravity model, in which the WALS language index is selected and weighted to measure the language distance and then introduced it into the trade gravity model as explanatory variable. The results show that: (1) language and bilateral trade flow are negatively related;(2) the language distance has negative significant influence on Thailand's bilateral trade flow;(3) language influence on Thailand's foreign trade cannot be overlooked. It should be highlighted by scholars as well as the relevant departments of the country;(4) the decline of global trade volume is seriously affected by the outbreak of COVID-19. © 2022 ACM.
ABSTRACT
This paper presents the results of electronic invoicing adoption and eCommerce turnovers in European Union enterprises between the years 2018 and 2020. Given the changes in e-invoicing adoption at national levels, the possible drivers of change are examined. As key explanatory variables are analyzed two sets of predictors. The first set is describing the overall embrace of ICT, digital skills, and innovations. The second set contains predictors describing the effects of changes related to the digitalization of enterprises. The results of the paper showed that enterprises from EU countries using e-invoicing on a smaller scale before 2018 are nowadays implementing more. Results showed that countries where at least one-third of enterprises are using e-invoicing also experienced higher growth in e-commerce turnovers, and cross-border sales, and the variability of eCommerce turnovers are lower in countries, where is a higher rate of e-invoicing adoption. © 2022 IDIMT. All rights reserved.
ABSTRACT
With the increased cost of living, exacerbated by COVID-19, thousands in New Jersey lack secure access to nutritious food. Given the importance of the issue, this research aims to produce an accurate metric using accessible data to quantify food insecurity. 16 potential explanatory variables, such as median household income and homeless population, were chosen as their values were defined across all NJ counties from 2015-2019. Using multiple linear regression, 14 unique metrics were created after four different variable pruning methods. The leading metric, with an adj. R2 value of 0.932, demonstrates the correlation between food insecurity, population, median household income, total population with health insurance, and population with private health insurance. The implementation of this metric could serve as a tool in predicting areas of food insecurity, highlighting affiliated factors, and revealing connections between racial populations. © 2021 IEEE.