A Framework for Learning Event Sequences and Explaining Detected Anomalies in a Smart Home Environment

Area of Science:

• Artificial Intelligence
• Smart Home Technology
• Cybersecurity

Background:

• Smart home systems generate vast amounts of data from Internet of Things (IoT) devices.
• Detecting anomalies in user behavior is crucial for security and system optimization.
• Existing methods often lack explainability for detected anomalies.

Purpose of the Study:

• To develop a framework for learning event sequences for anomaly detection in smart homes.
• To enhance the explainability of detected anomalies.
• To model and identify deviations from habitual user behavior.

Main Methods:

• Modeling user behavior as event sequences from IoT device interactions.
• Learning habitual behavior from recorded event sequences.
• Implementing anomaly detection as deviations from learned normal behavior.
• Utilizing simple rule analysis for anomaly explainability.

Main Results:

• A framework capable of learning user behavior patterns from smart home event data.
• Successful identification of anomalies as deviations from learned habitual behavior.
• Demonstrated explainability of detected anomalies through rule analysis.

Conclusions:

• The proposed framework effectively detects anomalies in smart home environments by learning event sequences.
• Explainability of anomalies is achieved, providing insights into deviations from normal user behavior.
• The method offers a robust approach for enhancing smart home security and performance.