Methods and strategies of precision marketing complex system for big data using recommendation algorithm

Abstract

This research aims to explore precision marketing methodologies and strategies within the domain of big data, with a specific focus on the recommendation algorithm. The initial phase involves an in-depth analysis of the system prerequisites for a user-centric personalized recommendation system rooted in big data. Following this, the study introduces the Latent Factor-Based Matrix Factorization Completion Based Hybrid Weighted Recommendation Method (LF-WMC). Moreover, considering the neighbor information set of customer and item, the above two prediction results are mixed to get a new prediction result according to the local and global impact of customer and item core. Finally, the weighted average is conducted based on the root mean squared error (RMSE) of the three prediction results. The LF-WMC proves to be a robust solution for addressing challenges like cold start and high data sparseness in big data precision marketing systems. It significantly enhances the accuracy of the system’s predictive recommendations. Experimental results demonstrate that, compared to other methods using the same dataset, LF-WMC consistently achieves lower root mean squared error (RMSE). Furthermore, LF-WMC exhibits superior accuracy, especially in scenarios with substantial data sparsity across multiple datasets.