The invention relates to a short-time passenger flow prediction method for a new line
station of
urban rail transit , which comprises the steps of S1, cleaning and preprocessing
station data, passenger flow data and POI data required by a prediction model, and dividing passenger flow into different stages after a new line is opened; S2, on the basis of the POI data cleaned in the S1, clustering point
land use functions of the whole website by taking the
land use scale around the new line
station as coordinates; S2, constructing a new line station historical passenger flow
database by combining the existing stations of which the clustering results are consistent with the new line station with the real-time new line passenger flow data obtained in the step S1; S3, based on S2, creating the historical passenger flow
database for the new line station; S4, converting the short-time passenger flow volume into a passenger
flow time-sharing coefficient based on the passenger flow data obtained in S2; and S5, combining the all-day passenger flow data in the S3 with the passenger
flow time-sharing coefficient in the S4 to obtain a short-time passenger flow prediction value result of the new line station. According to the invention, the passenger flow prediction precision is improved, and the service level of
urban rail operation is improved.