Digital Media Platform under TV Program Innovation Design and Dissemination Effect Evaluation

At present, the highly user experience and content communication effect evaluation of TV programs become more and more prominent with the fast-pace explosion of digital media platforms. Based on the existing multimodal fusion recommendation system and deep learning algorithm, this paper proposes an innovative model and constructs a multi-level feedback mechanism by fusing audience interaction data, social media feedback and TV program content characteristic. And it adds the sentiment analysis module and dissemination path tracing module to the conventional recommendation system, thus enabling more accurate assessment of the dissemination effect of the program and optimized personalized programming content design using reinforcement learning. The dynamic adjustment of program innovation design and effect evaluation is realized by predicting audience behavior and analyzing the propagation path that we use multi-layer long short-term memory network to achieve. Experimental results show that compared with the traditional method in the evaluation of communication effect of different types of TV programs, the accuracy of prediction of different TV program can be improved through the proposed model.