Hybrid CNN-GRU Framework for Intelligent Video Advertisement Classification

The fast growth of online advertising has made it even more necessary to have practical, automated video advertisement classification to make targeted marketing strategies more effective. This paper presents a Hybrid Deep Learning Framework for automated video advertisement classification, which combines Convolutional Neural Networks (CNNs) for learning spatial features and Gated Recurrent Units (GRUs) for learning sequential dependencies. The model leverages Natural Language Processing (NLP) techniques, including tokenization, removal of stopwords, and lemmatization, to preprocess video metadata to enhance the accuracy of ad classification. The supervised learning technique uses a labeled video advertisement dataset to extract text and metadata features and preprocess for more accurate ad classification. Experimental results demonstrate the effectiveness of the proposed CNN-GRU hybrid model over traditional deep learning frameworks in terms of classification accuracy, achieving 97.92% with enhanced generalizability across different advertisement categories. These outcomes demonstrate the ability of hybrid deep learning models to enable optimized digital marketing campaigns, increase content personalization, and simplify massive-scale advertisement classification. Future research will include attention mechanisms and transformer models to optimize classification efficiency and scalability