Research on New Media Marketing Strategies Based on Information Technology in the Big Data Era

Abstract

In the era of big data, the rapid development of information technology and the prevalence of the Internet have generated massive user behavior data on new media platforms. Extracting valuable information from this deluge of data and making precise recommendations based on user behavior characteristics and interest preferences have become crucial issues in new media marketing. To address this, this study proposes a content recommendation algorithm that integrates Graph Neural Network (GNN) and Self-Attention Mechanism, aiming to enhance the training efficiency and prediction accuracy of the model. Experimental results show that the proposed algorithm achieves an accuracy of 0.8 for new user recommendations, 0.85 for new item recommendations, with a recommendation precision of 0.88, a coverage rate of 0.75, and a diversity score of 0.7. Compared to collaborative filtering, content-based recommendation, deep neural network recommendation systems, and matrix factorization systems, the proposed algorithm demonstrates significant advantages across all metrics. Notably, it achieves accuracies of 0.82 and 0.84 for user cold-start and item cold-start problems, respectively. On the Amazon Reviews dataset, the electronics category excels in user cold-start accuracy (0.82), item cold-start accuracy (0.84), recommendation precision (0.88), coverage (0.73), and diversity (0.68). Additionally, favorable results are observed in the food and book categories. Overall, the proposed GNN and Self-Attention Mechanism-based recommendation algorithm significantly improves the accuracy and efficiency of user behavior prediction and content recommendation.