Traffic police level prediction method based on distance metric learning

Abstract

A traffic police level prediction method based on distance metric learning, which predicts the urban traffic police level under known weather data, time data, environmental data, etc., and organizes multi-dimensional historical data according to the requirements of the traffic police command department Classification, using the generalized Mahalanobis distance measurement method to learn the multi-dimensional historical data after classification and marking, the distance measurement learning matrix obtains the weight of each feature attribute to the traffic police level, and the feature attribute with a large weight contributes a lot to the classification , calculate the similarity between the current multi-dimensional data and historical data according to the weighted Euclidean distance, select K historical data most similar to the current data to vote on the police level, and the police level with the highest vote is the current traffic police level prediction results. The invention effectively realizes prediction and has better accuracy.

Description

technical field [0001] The invention belongs to the field of intelligent traffic, and in particular relates to a method for predicting urban traffic police levels. Background technique [0002] With the rapid development of the economy, the rapid growth of the number of motor vehicles in the urban traffic system has greatly increased the probability of traffic accidents and traffic congestion. For the motor vehicle drivers on the road, the traffic management personnel hope to obtain the traffic police situation level in the macro region. An effective prediction of the regional police situation level in a certain period of time in the future will help the traffic management department to optimize the deployment of police force and formulate corresponding measures. plan to ease the traffic pressure in key areas. [0003] Patent 201410610003.7 collects traffic flow data including working days, non-working days and major holidays, reorganizes the traffic flow data in the same c...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the disadvantages of no prediction function and poor accuracy in the existing traffic police situation discrimination methods, the present invention provides a traffic police situation level prediction method based on distance metric learning that can effectively realize prediction and have good accuracy.

Images