Improvement of the CNN - LSTM Deep Neural Network Model Algorithm for Personalized Movie - watching Recommendation

With the growing demand for personalized movie - watching among users of streaming platforms, this paper proposes an improvement to the algorithm based on the CNN - LSTM deep neural network model. This improvement is used to extract text sequence information for movie - watching recommendations. Specifically, the CNN model extracts text features and processes these features using the convolutional network of CNN. By combining the temporal sequence modeling and interest evolution capabilities of LSTM, it can effectively uncover the long - term interest preferences in users' movie - watching records, ratings, and reviews. Through dynamically adjusting interest weights and thresholds, this algorithm optimizes the recommendation accuracy, alleviates the problems of data sparsity and interest drift, and provides a more personalized movie - watching - aware recommendation solution for video platforms.