An Efficient Recommendation Filter Model on Smart Home Big Data Analytics for Enhanced Living Environments.

Area of Science:

• Smart Home Technology
• Recommender Systems
• Human-Computer Interaction

Background:

• Smart home user interfaces are complex and inflexible, demanding significant user adaptation.
• Existing systems struggle to intuitively predict user needs in enhanced living environments.

Purpose of the Study:

• To develop a weighted hybrid recommender system for predicting user intentions in smart homes.
• To enhance user experience by simplifying interaction with complex smart home systems.

Main Methods:

• A weighted hybrid recommender system integrating contextual collaborative filtering and content-based recommendations.
• Implementation of an adaptive Kalman Filter model for dynamic weight adjustment of system components.
• Weight hybridization method to combine different recommendation strategies optimally.

Main Results:

• The proposed system dynamically predicts and revises component weights for optimal performance.
• Experimental results demonstrate improved recall and precision rates compared to existing methods.
• The system effectively optimizes weight distribution among its components.

Conclusions:

• The weighted hybrid recommender system significantly enhances smart home user experience.
• The adaptive Kalman Filter model provides stability and optimizes prediction accuracy.
• This approach offers a more intuitive and efficient way for users to interact with smart homes.