Comprehensive traffic data mining method based on multi-source data fusion

Abstract

The invention discloses a comprehensive traffic data mining method based on multi-source data fusion and belongs to the technical field of data processing. The data mining method comprises the following specific steps of (1) collecting the traffic information; (2) data preprocessing; (3) feature extraction; (4) constructing an initial prediction model; (5) fusing decision-level data; (6) constructing a fusion traffic prediction model; (7) evaluating the model; and (8) applying the model; compared with a traditional single data acquisition mode, the multi-source data sensor is adopted for comprehensive traffic data analysis and processing, a fused traffic prediction model is constructed for traffic intersection congestion condition prediction, prediction precision of the traffic congestion condition can be improved, therefore, the traffic department can carry out traffic dispersion through the traffic information issuing platform.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a comprehensive traffic data mining method for multi-source data fusion. Background technique [0002] After retrieval, Chinese Patent No. CN103700255A discloses a traffic flow prediction method based on spatio-temporal correlation data mining. Although the structure of the invention is simple, its data acquisition channel is single, resulting in limited prediction accuracy; data fusion is a multi-level and multi-faceted process. The basic principle of the data processing process is to make full use of the resources of each sensor, through the rational use of the observation information obtained by these sensors, and to combine a variety of redundant and complementary sensors in space or time according to a certain principle, so as to obtain The consistent description of the measured object improves the effectiveness of the sensor; the essence of multi-sensor information ...

AI Technical Summary

Problems solved by technology

[0002] After retrieval, Chinese Patent No. CN103700255A discloses a traffic flow prediction method based on spatio-temporal correlation data mining. Although the structure of the invention is simple, its data acquisition channel is single, resulting in limited prediction accuracy; data fusion is a multi-level and multi-faceted process. The basic principle of the data processing process is to make full use of the resources of each sensor, through the rational use of the observation information obtained by these sensors, and to combine a variety of redundant and complementary sensors in space or time according to a certain principle, so as to obtain The consistent description of the measured object improves the effectiveness of the sensor; the essence of multi-sensor information fusion is that it can appear in different dimensions, different levels, and different time periods, has a more complex nature, and is closer to human The intelligent calculation of the brain; nowadays, with the rapid economic development and the increasing number of private vehicles, urban congestion is becoming more and more serious; and the current traffic agencies mostly use a single channel to obtain traffic information, and use this as the basis for traffic forecasting For predicting traffic congestion, due to the single channel for obtaining traffic information, it is easy to reduce the accuracy of traffic forecasting, which is not conducive to traffic guidance; therefore, it is particularly important to invent a comprehensive traffic data mining method for multi-source data fusion;

[0003] Most of the existing comprehensive traffic data mining methods obtain the characteristics of traffic data by extracting a certain data holder (such as the traffic bureau or traffic APP), and analyze and process the traffic flow data according to it. The single source of data acquisition leads to limited prediction precision, which is not conducive to traffic guidance; therefore, we propose a comprehensive traffic data mining method based on multi-source data fusion

Images