Research on Optimization of New Media Marketing Strategies Based on Machine Learning Algorithms

As Internet technologies advance and consumer expectations for personalized experiences rise, there is an increasing need for optimized marketing strategies that can effectively engage audiences. With the proliferation of digital platforms, there is a pressing need to enhance marketing accuracy and consumer targeting. Therefore, the optimization of the new media marketing strategy is performed by implementing a novel Sooty Tern Optimized Efficient Decision Tree (STO-EDT). The data used in this research comprise interaction with posts such as likes, shares, comments, clicks, engagement with ads, time spent on platforms, and purchase history. Further, preprocessing involves handling missing values and normalization to scale numeric data into a specific range. Next, feature extraction is involved, which applies the Term Frequency-Inverse Document Frequency (TF-IDF) approach to extract important keywords from text data. The hybrid approach STO-EDT is executed to predict customer behavior, classify engagement patterns, and optimize marketing decisions. Experimental results demonstrate that the STO-EDT algorithm outperforms traditional approaches in terms of recall (97.1%), accuracy (97.4%), F1-score (98.3%), and precision (98.2%). The optimized decision tree structure offers better insights into customer behavior, leading to more effective marketing strategies.