Commodity Recommendation Model Integrating User Psychological Data Analysis

Area of Science:

• E-commerce
• Consumer Psychology
• Recommendation Systems

Background:

• E-commerce recommendation systems are crucial for managing product information overload and enhancing user experience.
• Existing methods include correlation-based (e.g., collaborative filtering) and causality-based recommendations.
• Causality-based methods, incorporating consumer psychology, excel in new product recommendations.

Purpose of the Study:

• To explore product recommendation using multiple consumer psychologies: motivation, attitude, and interest.
• To analyze recommendation patterns and accuracy based on single, dual, and triple psychological factors.

Main Methods:

• Selected product-store attributes based on consumption psychology theory and e-commerce needs.
• Conducted a comprehensive comparison of multiple consumer psychology patterns in product recommendations.
• Analyzed recommendation stability and result patterns, comparing single, dual, and triple psychology approaches.

Main Results:

• Dual and triple consumer psychology recommendation methods effectively improve accuracy over single methods.
• Dual consumer psychology recommendations generally show higher accuracy than triple psychology methods.
• The highest recommendation accuracy was achieved using a dual approach combining consumer motivation and attitude psychology.

Conclusions:

• Incorporating multiple consumer psychologies significantly enhances e-commerce product recommendation accuracy.
• A dual-psychology model, particularly motivation-attitude, offers superior performance compared to single or triple models.
• Causality-based recommendations leveraging consumer psychology provide a valuable alternative to traditional correlation-based methods.