도로 교통량 조사 데이터 특성 분석 및 결측 보정 방법 (A Study on the Analysis of Road Traffic Survey Data Characteristics and Missing Correction Method)

서지정보

· 발행기관 : 한국도로학회
· 수록지 정보 : 한국도로학회논문집 / 25권 / 5호 / 205 ~ 211페이지
· 저자명 : 한대철, 문학룡, 변상철, 이상수

초록

PURPOSES : Traffic volume, an important basic data in the field of road traffic, is collected from traffic survey equipment installed at certain locations, which sometimes results in missing traffic volume data and abnormal detection. Therefore, this study presents various missing correction techniques using traffic characteristic analysis to obtain accurate traffic volume statistics.
METHODS : The fundamental premise behind the development of a traffic volume correction and prediction model is to set the corrected data as the reference value, and the traffic volume correction and prediction process for the outliers and missing values in the raw data were performed based on the set values.
RESULTS : The simulation results confirmed that the algorithm combining seasonal composition, quantile AD, and aggregation techniques showed a detection performance of more than 91% compared with actual values.
CONCLUSIONS : Raw data collected due to difficulties faced by traffic survey equipment will result in missing traffic volume data and abnormal detection. If these abnormal data are used without appropriate corrections, it is difficult to accurately predict traffic demand.
Therefore, it is necessary to improve the accuracy of demand prediction through characteristic analysis and the correction of missing data or outliers in the traffic data.

영어초록

PURPOSES : Traffic volume, an important basic data in the field of road traffic, is collected from traffic survey equipment installed at certain locations, which sometimes results in missing traffic volume data and abnormal detection. Therefore, this study presents various missing correction techniques using traffic characteristic analysis to obtain accurate traffic volume statistics.
METHODS : The fundamental premise behind the development of a traffic volume correction and prediction model is to set the corrected data as the reference value, and the traffic volume correction and prediction process for the outliers and missing values in the raw data were performed based on the set values.
RESULTS : The simulation results confirmed that the algorithm combining seasonal composition, quantile AD, and aggregation techniques showed a detection performance of more than 91% compared with actual values.
CONCLUSIONS : Raw data collected due to difficulties faced by traffic survey equipment will result in missing traffic volume data and abnormal detection. If these abnormal data are used without appropriate corrections, it is difficult to accurately predict traffic demand.
Therefore, it is necessary to improve the accuracy of demand prediction through characteristic analysis and the correction of missing data or outliers in the traffic data.

참고자료

· 없음