Driving behavior scoring method based on logistic regression and single-level analysis weighting method

Abstract

The invention relates to a driving behavior scoring method based on logistic regression and a single-level analysis weighting method. The driving behavior scoring method comprises the following steps:S1, collecting driving behavior data; s2, performing abnormal value elimination and data preprocessing on the driving behavior data; s3, determining a driving behavior scoring index system; s4, constructing a logistic regression model to determine the relative importance degree of each scoring index; s5, determining the weight of the driving behavior scoring index through a single-level analysisweighting method; and S6, calculating a driving behavior score according to a scoring system formula. According to the driving behavior scoring method based on logistic regression and a single-level analysis weighting method, the problems that in the process of constructing a logistic regression model and extracting regression coefficients, the number of characteristic variables is large, and thedata size is insufficient are solved. In addition, the attention degree of the subjective and objective weighting method to different attribute indexes is balanced, a driving scoring model based on deep learning is provided, and by establishing the driving behavior scoring model, accurate and objective behavior scores can be provided for real-time driving behaviors.

Description

technical field [0001] The invention relates to the field of intelligent evaluation of driving behavior, and more specifically, relates to a driving behavior scoring method based on logistic regression and single-level analysis weighting method. Background technique [0002] With the improvement of living standards, the number of motor vehicles is also increasing. In the process of driving, unsafe driving behavior has become the source of potential safety hazards. According to statistics, bad driving behaviors account for 86.5% of traffic accident deaths. Therefore, it is imminent to judge the quality of driving behavior qualitatively and quantitatively. [0003] In the process of driving behavior discrimination, due to the differences in the importance of each index, the work of index weighting becomes very important. At present, the mainstream empowerment methods are divided into objective empowerment method and subjective empowerment method. However, there are certain li...

AI Technical Summary

Problems solved by technology

At present, the mainstream empowerment methods are divided into objective empowerment method and subjective empowerment method. However, there are certain limitations when using any of the above methods alone:

[0004] (1) Objective weighting methods such as principal component analysis, entropy technology, mean square error, and goal programming are a type of method in which each index is automatically weighted according to certain rules. The sample data and actual problems are poor in versatility and participation;

[0005] (2) Subjective weighting methods such as expert survey method, binomial coefficient method, ring-comparison scoring method and analytic hierarchy process are greatly affected by the subjective importance of decision makers. Due to differences in professional level, functions and business fields, opinions differ Larger, it is difficult to obtain a unified result, which reduces the credibility of expert opinions to a certain extent

[0006] Analytic Hierarchy Process is a system analysis method that combines qualitative and quantitative aspects of non-quantitative events in systems engineering. However, in actual project evaluation, it is difficult to obtain the relative importance of indicators more accurately.

Images