Deep Learning Versus Traditional Solutions for Group Trajectory Outliers.

Area of Science:

• Computer Science
• Data Science
• Artificial Intelligence

Background:

• Trajectory data analysis is crucial for understanding movement patterns.
• Identifying outliers in large trajectory datasets presents significant computational challenges.
• Existing methods often struggle with accuracy and efficiency.

Purpose of the Study:

• To introduce a novel model for identifying groups of trajectory outliers.
• To propose and evaluate multiple algorithms for outlier detection.
• To compare the performance of different algorithmic approaches.

Main Methods:

• Development of algorithms based on data mining and knowledge discovery.
• Implementation of machine learning and computational intelligence techniques, including ensemble learning and metaheuristics.
• Exploration of a convolution deep neural network for feature learning and outlier identification.

Main Results:

• Experimental validation on diverse trajectory databases was conducted.
• The deep learning approach demonstrated superior performance compared to other methods.
• The proposed convolution deep neural network achieved state-of-the-art runtime and accuracy.

Conclusions:

• Deep learning offers a highly effective solution for group trajectory outlier detection.
• The novel convolution deep neural network model provides significant advantages in performance.
• This research advances the field of trajectory data analysis and outlier identification.