Night stay parking demand predicting method based on survival analysis

Abstract

The invention discloses a night stay parking demand predicting method based on survival analysis. The night stay parking demand predicting method includes the steps that parking lot continuous parkingdata and some external factor data are obtained; the day time period for parking and the night time period for parking are divided according to parking feature analysis, and night parking demands areconverted into a stay part for day vehicle parking; a parking event is expressed with a survival event, corresponding data processing is conducted, and thus the form of the parking event is applied to a survival analysis method; a semi-parametric model method is adopted to establish a Cox proportional risk model with multiple factors affecting the parking time; and according to a model result, the probability of different parking time of the day driven-in vehicles under different conditions is predicted, and predicted night parking demands are further obtained. According to the new night parking demand predicting microcosmic method, good prediction precise is achieved. According to the night stay parking demand predicting method, the basis can be provided for night demand management of aparking lot, and the night stay parking demand predicting method can be used for the aspects such as pricing, shared parking space opening, and parking partition in fine management.

Description

technical field [0001] The invention relates to the field of static traffic and parking lot design, in particular to a method for forecasting parking demand at night based on survival analysis. Background technique [0002] With the continuous increase of the number of motor vehicles, insufficient parking facilities have become a common problem in major cities, and refined parking management has become an important means to improve the utilization efficiency of parking facilities and alleviate parking difficulties. Premise and basis, it is very important to accurately predict parking demand. With the deepening of the concept of shared parking, refined parking management is bound to distinguish between day and night parking needs according to the main building type of the parking lot, and formulate corresponding and reasonable parking space reservation capacity plans, so as to achieve full parking demand Grasp the daily changes and ensure the orderly implementation of shared...

AI Technical Summary

Problems solved by technology

[0004] Although the application of artificial intelligence methods such as neural networks can theoretically obtain better prediction accuracy, because this type of method contains some unrecognizable (implicit) process methods, researchers cannot further grasp the changing law of parking demand. At the same time, it also greatly reduces the guiding significance of the prediction results for parking management decisions.

Images