[학위논문]Transit Ridership Models Using Smart Card Big Data : The Two-stage Least Squares Method(2SLS) and The Generalized Method of Moments(GMM) Approaches원문보기
본 연구는 지하철 이용자 수요예측을 위하여 내·외생적 영향요인들을 분석하고, 각 요인들이 일반선형회귀모형 상에서 내재하고 있는 편의를 통제하여 보다 설명력이 높은 예측 방법론을 제시하는데 주요 목적이 있다. 지하철 이용수요와 관련하여 기존 연구들은 주로 사회경제 및 토지이용 등의 요인들과의 상관관계 위주의 연구를 수행해 왔으나, 버스 등의 타 대중교통 수요의 내생성 및 지하철역 주변상업용도의 외생적 영향에 대한 연구는 고려하지 않았다. 따라서 본 연구는 이러한 문제들을 개선하기 위하여 ...
본 연구는 지하철 이용자 수요예측을 위하여 내·외생적 영향요인들을 분석하고, 각 요인들이 일반선형회귀모형 상에서 내재하고 있는 편의를 통제하여 보다 설명력이 높은 예측 방법론을 제시하는데 주요 목적이 있다. 지하철 이용수요와 관련하여 기존 연구들은 주로 사회경제 및 토지이용 등의 요인들과의 상관관계 위주의 연구를 수행해 왔으나, 버스 등의 타 대중교통 수요의 내생성 및 지하철역 주변상업용도의 외생적 영향에 대한 연구는 고려하지 않았다. 따라서 본 연구는 이러한 문제들을 개선하기 위하여 2단계 최소자승법(2SLS)과 일반적률추정법(GMM)을 적용하였으며, 연구결과를 통하여 버스 이용자수와 상업용도 요인들의 내·외생성은 2SLS 추정에 의해 지하철 이용수요 예측을 보다 고도화 할 수 있는 것으로 나타났다.
본 연구는 지하철 이용자 수요예측을 위하여 내·외생적 영향요인들을 분석하고, 각 요인들이 일반선형회귀모형 상에서 내재하고 있는 편의를 통제하여 보다 설명력이 높은 예측 방법론을 제시하는데 주요 목적이 있다. 지하철 이용수요와 관련하여 기존 연구들은 주로 사회경제 및 토지이용 등의 요인들과의 상관관계 위주의 연구를 수행해 왔으나, 버스 등의 타 대중교통 수요의 내생성 및 지하철역 주변상업용도의 외생적 영향에 대한 연구는 고려하지 않았다. 따라서 본 연구는 이러한 문제들을 개선하기 위하여 2단계 최소자승법(2SLS)과 일반적률추정법(GMM)을 적용하였으며, 연구결과를 통하여 버스 이용자수와 상업용도 요인들의 내·외생성은 2SLS 추정에 의해 지하철 이용수요 예측을 보다 고도화 할 수 있는 것으로 나타났다.
With the rapid growth of the global economy, the ownership of cars is increasing, at the same time, environment pollution and traffic congestion problems becoming more and more severe. This is against people’s health and wasted lots of time and money. To build a sustainable society, advocating the w...
With the rapid growth of the global economy, the ownership of cars is increasing, at the same time, environment pollution and traffic congestion problems becoming more and more severe. This is against people’s health and wasted lots of time and money. To build a sustainable society, advocating the widespread transit use is a well-known measure. However, it is difficult to develop strategies without knowing the factors affecting transit ridership. The objective of this study is to develop a consistent model for predicting subway ridership. Previous studies have focused on the relationship between subway ridership and factors such as socio-economy and land-use. However, there was a lack of research on the endogenous and heterosecdasticity impacts of alternative transit demand and exogenous trade area. Therefore, this study incorporated the effects of bus ridership and retail properties within 300m and 1,000 meters radius from 248 subway stations in Seoul in the modeling process. We also selected seven factors(bus stop, bus line, sulfurous acid gas ppm, carbon monoxide ppm, nitrogen monoxide, fine dust, ozone ppm) as the instrumental variables to estimate bus ridership which was endogenous. In a methodological aspect, we applied two-stage least squares Method(2SLS) approach to capturing the endogeneity issue and the generalized method of moments(GMM) approach to capturing the heteroscedasticity issue. Compared with other studies, there have three differences. The first difference is separating the transit ridership into boarding and deboarding. The second difference is that one-day ridership is divided into four time periods, morning, afternoon, evening and dawn. The third difference is land use variables are the proportion of 10 retail types throughout Seoul city around each subway station within 300 meters and 1,000 meters radius separately. As a result, bus ridership and attributes for trade area were found to be significant, and the methodological approach for the model development was plausible to estimate subway ridership. Based on these findings, strategies could be proposed to improving the substantiality activities and eco-friendly transportation when the new bus and subway routes are planned.
With the rapid growth of the global economy, the ownership of cars is increasing, at the same time, environment pollution and traffic congestion problems becoming more and more severe. This is against people’s health and wasted lots of time and money. To build a sustainable society, advocating the widespread transit use is a well-known measure. However, it is difficult to develop strategies without knowing the factors affecting transit ridership. The objective of this study is to develop a consistent model for predicting subway ridership. Previous studies have focused on the relationship between subway ridership and factors such as socio-economy and land-use. However, there was a lack of research on the endogenous and heterosecdasticity impacts of alternative transit demand and exogenous trade area. Therefore, this study incorporated the effects of bus ridership and retail properties within 300m and 1,000 meters radius from 248 subway stations in Seoul in the modeling process. We also selected seven factors(bus stop, bus line, sulfurous acid gas ppm, carbon monoxide ppm, nitrogen monoxide, fine dust, ozone ppm) as the instrumental variables to estimate bus ridership which was endogenous. In a methodological aspect, we applied two-stage least squares Method(2SLS) approach to capturing the endogeneity issue and the generalized method of moments(GMM) approach to capturing the heteroscedasticity issue. Compared with other studies, there have three differences. The first difference is separating the transit ridership into boarding and deboarding. The second difference is that one-day ridership is divided into four time periods, morning, afternoon, evening and dawn. The third difference is land use variables are the proportion of 10 retail types throughout Seoul city around each subway station within 300 meters and 1,000 meters radius separately. As a result, bus ridership and attributes for trade area were found to be significant, and the methodological approach for the model development was plausible to estimate subway ridership. Based on these findings, strategies could be proposed to improving the substantiality activities and eco-friendly transportation when the new bus and subway routes are planned.
AI-Helper
※ AI-Helper는 부적절한 답변을 할 수 있습니다.