Constructing Trajectory Trait Profile Of Private Vehicle Driver Based On The Big Five Personality

The accumulation of a large amount of mobile trajectory data has made it possible to mine individual travel behavior patterns.Establishing driver profiles based on private vehicle trajectory not only contributes to driving safety but also guides personalized location-based service recommendations and intelligent transportation construction.However,existing methods for constructing driver profiles often only describe driving behavior based on vehicle driving features such as speed and acceleration,lacking characterization of individual spatiotemporal mobility patterns.On the other hand,research on mining individual mobility patterns mostly relies on low-level trajectory features of spatiotemporal or geographic semantics to interpret mobility patterns,predict travel behavior,or infer identity information.Few studies consider multidimensional trajectory features and capture stable traits that reflect travel tendencies and driving preferences to establish driver profiles,which restricts the application scenarios of trajectory analysis.Therefore,in this study,drawing on the Big Five Personality theory,we propose the trajectory trait profile to achieve a comprehensive and concise characterization of driver mobility patterns and driving behavior.The trajectory trait profile includes four traits: trajectory extraversion,trajectory openness,trajectory neuroticism,and trajectory conscientiousness.Around the measurement of trajectory traits,this research includes three parts: the development and evaluation of the Trajectory Trait Scale(TTS),and the analysis of measurement results.1)Development of Trajectory Trait Scale: The TTS serves as the mapping and transformation mechanism between the underlying trajectory features and the abstract trajectory traits.The development of this scale involves extracting trajectory features from four perspectives: time,space,semantics,and driving behavior,and assigning appropriate trajectory features as items for each sub-scale.2)Evaluation of Trajectory Trait Scale: Evaluating the stability and validity of measurement results using the TTS is a prerequisite for applying the scale.In this study,we consider four evaluation indicators,including item discrimination,internal consistency,split-half reliability,and validity.3)Analysis of Trajectory Trait Profiles: We discuss the characteristics of trajectory trait profile from the perspectives of trajectory features,trajectory traits,and the stability of trajectory profiles to deepen the understanding of trajectory traits.Specifically,we explore the distribution patterns of trajectory features and the correlations among features.Besides,we analyze the correlations among trajectory traits and identify the typical driver types using clustering algorithms.The impacts of trajectory completeness,seasonal changes,and traffic conditions on the stability of trajectory trait profiles are also discussed.Based on trajectory data from 662 anonymous private car drivers in Shenzhen over an 8-month period,experiments of scale evaluation and profile analysis are conducted.The experimental results show that all 32 items in the TTS have significant item discrimination,and the measurement content of the items within the four sub-scales is consistent.The measurement results are stable and aligned with the definition of trajectory traits,ensuring the validity of the scale.In addition,trajectory integrity,seasonal changes,and traffic performance have minimal impact on the stability of trajectory profiles.This result can guide data collection and selection for profiling trajectory and provide a reference for profile reliability.Finally,this paper discusses the application scenarios of trajectory trait profiles and demonstrates the potential value of profiles with the example of usage-based insurance premium determination.