Anomaly detection and restoration method for multi-dimensional highway toll collection data

Abstract

The invention discloses an anomaly detection and restoration method for multidimensional highway toll collection data, and the method comprises the following steps: 1, collecting the original multidimensional toll collection data of a highway, and carrying out the preprocessing of the original multidimensional toll collection data of the highway, so as to obtain original data after multidimensional standardization; 2, constructing an abnormal data detection model based on the sum of similarity coefficients; 3, inputting the original data after the multi-dimensional specification obtained in the step 1 into an abnormal data detection model based on a similarity coefficient sum, and obtaining multi-dimensional data containing an abnormal value through detection; 4, constructing an abnormal data restoration model based on extreme gradient promotion; and step 5, inputting the multi-dimensional data containing the abnormal value obtained in the step 3 into an abnormal data restoration modelbased on extreme gradient promotion to realize multi-dimensional data anomaly restoration and effect evaluation. According to the invention, the quality and availability of highway toll collection data can be obviously improved, and a good data basis is provided for subsequent highway abnormal event detection and big data statistical analysis work.

Description

technical field [0001] The invention belongs to the field of data mining, and discloses a method for abnormality detection and repair of multi-dimensional expressway toll data. Background technique [0002] With the construction of the expressway network and the advent of the information age, the intelligent toll collection system is becoming more and more perfect, and the collected data has reached a considerable level. Among them, the expressway toll data that occupies an important position has detailed vehicle traffic information, and data mining technology can solve expressway operation and decision-making problems, but this needs to be based on high-quality data. There are often "dirty data" represented by outliers and missing values ​​in the charging data, which greatly affects the effect of data mining. If the single-dimensional abnormal data cleaning method is directly used and the "dirty data" is simply filtered out, a large amount of attributes and information wil...

AI Technical Summary

Problems solved by technology

There are often "dirty data" represented by outliers and missing values ​​in charging data, which greatly affects the effect of data mining

Images