Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Neural Collaborative Filtering to Detect Anomalies in Human Semantic Trajectories

Yueyang Liu, Lance Kennedy, Hossein Amiri, Andreas Züfle

TL;DR

The proposed Neural Collaborative Filtering approach is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations.

Abstract

Human trajectory anomaly detection has become increasingly important across a wide range of applications, including security surveillance and public health. However, existing trajectory anomaly detection methods are primarily focused on vehicle-level traffic, while human-level trajectory anomaly detection remains under-explored. Since human trajectory data is often very sparse, machine learning methods have become the preferred approach for identifying complex patterns. However, concerns regarding potential biases and the robustness of these models have intensified the demand for more transparent and explainable alternatives. In response to these challenges, our research focuses on developing a lightweight anomaly detection model specifically designed to detect anomalies in human trajectories. We propose a Neural Collaborative Filtering approach to model and predict normal mobility. Our method is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations. Our algorithm consists of two main modules. The first is the collaborative filtering module, which applies collaborative filtering to model normal mobility of individual humans to places of interest. The second is the neural module, responsible for interpreting the complex spatio-temporal relationships inherent in human trajectory data. To validate our approach, we conducted extensive experiments using simulated and real-world datasets comparing to numerous state-of-the-art trajectory anomaly detection approaches.

Neural Collaborative Filtering to Detect Anomalies in Human Semantic Trajectories

TL;DR

The proposed Neural Collaborative Filtering approach is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations.

Abstract

Human trajectory anomaly detection has become increasingly important across a wide range of applications, including security surveillance and public health. However, existing trajectory anomaly detection methods are primarily focused on vehicle-level traffic, while human-level trajectory anomaly detection remains under-explored. Since human trajectory data is often very sparse, machine learning methods have become the preferred approach for identifying complex patterns. However, concerns regarding potential biases and the robustness of these models have intensified the demand for more transparent and explainable alternatives. In response to these challenges, our research focuses on developing a lightweight anomaly detection model specifically designed to detect anomalies in human trajectories. We propose a Neural Collaborative Filtering approach to model and predict normal mobility. Our method is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations. Our algorithm consists of two main modules. The first is the collaborative filtering module, which applies collaborative filtering to model normal mobility of individual humans to places of interest. The second is the neural module, responsible for interpreting the complex spatio-temporal relationships inherent in human trajectory data. To validate our approach, we conducted extensive experiments using simulated and real-world datasets comparing to numerous state-of-the-art trajectory anomaly detection approaches.
Paper Structure (25 sections, 16 equations, 6 figures, 3 tables)

This paper contains 25 sections, 16 equations, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Trajectory Anomaly Detection
  4. Statistical Approaches:
  5. Knowledge-Driven Approaches
  6. Machine Learning Approaches
  7. Collaborative Filtering
  8. Preliminaries
  9. Human Semantic Trajectory Representation
  10. Human Spatial-temporal Trajectory Anomaly Detection
  11. Methodology
  12. Neural Collaborative Filtering Model
  13. Quantification of Anomaly Scores
  14. User-POI Matrix Reconstruction Surprise:
  15. POI Type Surprise
  16. ...and 10 more sections

Figures (6)

  • Figure 1: A sample of anomaly trajectory (red) and Normal trajectory (blue).
  • Figure 2: A stylized example of a User-POI matrix
  • Figure 3: Expected User-POI visits obtained from the User-POI Matrix of Figure \ref{['fig:upmatrix']}. Obtained via Randomized Singular Value Decomposition having three latent components.
  • Figure 4: Measuring Surprise between the expected User-POI Matrix and observed User-POI visits.
  • Figure 5: The illustration of our proposed NCF model. (Left) Extraction of spatio-temporal embeddings from input trajectories; (Middle) Neural Collaborative Filtering with two-level of prediction; (Right) Anomaly scores based on weighted prediction.
  • ...and 1 more figures

Theorems & Definitions (2)

  • definition 1: Train and Test Period
  • definition 2: Trajectory Anomaly Detection