An Integrated Fuzzy C-Means Method for Missing Data Imputation Using Taxi GPS Data

Area of Science:

• Intelligent Transportation Systems (ITS)
• Data Science
• Artificial Intelligence

Background:

• Traffic-sensing technologies are crucial for traffic control.
• Missing data in Intelligent Transportation System (ITS) sensing datasets, caused by device malfunctions or errors, degrade data quality.
• Accurate data imputation is essential for reliable traffic management.

Purpose of the Study:

• To propose an integrated imputation algorithm to enhance the accuracy of missing value estimation in ITS sensing datasets.
• To optimize the Fuzzy C-Means (FCM) model parameters using the Genetic Algorithm (GA).
• To validate the effectiveness of the proposed method using real-world taxi GPS data.

Main Methods:

• An integrated imputation algorithm combining Fuzzy C-Means (FCM) and Genetic Algorithm (GA).
• GA optimizes FCM parameters, including membership degree and number of cluster centroids.
• Experimental validation using taxi Global Positioning System (GPS) data from Manhattan, NYC.

Main Results:

• The integrated imputation method demonstrated superior performance compared to history imputation and conventional FCM.
• Achieved high Relative Accuracy (RA) values of 0.576 (±5%) and 0.785 (±10%).
• Euclidean distance yielded better imputation performance than Manhattan distance within the clustering approach.

Conclusions:

• The proposed integrated imputation method effectively estimates missing values in daily traffic management.
• The GA-optimized FCM approach offers a robust solution for improving ITS data quality.
• The findings support the use of advanced imputation techniques for reliable intelligent transportation systems.