Efficient and accurate personalized product recommendations through frequent item set mining fusion algorithm.

Area of Science:

• E-commerce
• Data Mining
• Recommender Systems

Background:

• Personalized product recommendation systems often struggle with sparse data and the cold start problem.
• Filtering invalid information in large e-commerce datasets is a significant, underexplored challenge.
• Existing fusion recommendation algorithms based on frequent item set mining face issues like redundant rules and low accuracy.

Purpose of the Study:

• To propose a novel fusion recommendation algorithm to address the challenge of filtering invalid information in e-commerce.
• To enhance the accuracy and efficiency of personalized product recommendations.
• To adapt to users' dynamic preferences and capture changing interests in real-time.

Main Methods:

• Developed a fusion recommendation algorithm based on frequent item set mining.
• Implemented data set compression and identification of frequent commodity sets.
• Incorporated calculation of user-commodity interest rankings and definition of similar product recommendation rules.

Main Results:

• The proposed algorithm effectively filters invalid commodity data, improving recommendation quality.
• Demonstrated improved time efficiency by reducing candidate frequent item sets.
• Achieved more accurate recommendations through user-commodity interest analysis and similar product identification.

Conclusions:

• The algorithm effectively filters e-commerce data, leading to more accurate and efficient personalized recommendations.
• It adapts to dynamic user preferences, enhancing user experience on e-commerce platforms.
• Comparative analysis shows reduced data sets and operation time compared to other data mining algorithms.