Analysis and Implementation of Human Mobility Behavior Using Similarity Analysis Based on Co-Occurrence Matrix.

Area of Science:

• Computer Science
• Data Science
• Human Mobility Analysis

Background:

• Information and Communications Technology (ICT) expansion provides extensive data for human mobility pattern analysis.
• Existing behavior-aware protocols aim to model mobile user characteristics and aggregation.
• Integrating physical and cyber-based systems requires advanced mobility analysis.

Purpose of the Study:

• To develop a similarity analytical framework for mobile encountering analysis.
• To propose a method for identifying similar user mobility patterns based on location and time.
• To create a technique for producing co-occurrence matrices to determine user encounters based on similar behaviors.

Main Methods:

• The Similarity Analysis Approach (SAA) utilizes device information such as Internet Protocol (IP) and Media Access Control (MAC) addresses.
• Analysis of user similarity distributions across different days and locations based on real movement data.
• Development of co-occurrence matrices to quantify user encounters derived from behavioral similarities.

Main Results:

• The SAA method demonstrates effective identification of users with common mobility behaviors.
• Analysis revealed similar user characteristics based on location and time, validating the approach's efficacy.
• The proposed SAA approach achieved a 33% higher accuracy in recognizing user similarity compared to existing methods.

Conclusions:

• The SAA provides a robust framework for integrating physical and cyber-based systems through mobility pattern analysis.
• The method accurately identifies and quantifies user encounters based on spatio-temporal behavioral similarities.
• SAA offers a significant improvement in the accuracy of human mobility pattern similarity recognition.