A Push Model of Advertisement Classification Matching Based on Machine Learning
Yan Sun, Qian Wu, Wendi Li · IOP Conference Series Materials Science and Engineering · 2020
Abstract With the leap-forward development of Internet technology, it is more effective for marketing to meet personalized needs based on massive data. In this paper, through the screening of advertisements and user feature vectors, combining factor decomposition machine and neural network, a multi-label model is built, and through training data, a static “advertisement-user” classification matching push model is obtained. On the basis of the static model, three sub-models are respectively established for comprehensive evaluation to realize the “user-advertisement” household matching push. At the same time, the recommendation model is updated in real time to establish the “channel user-video advertisement” classification matching push update model, and the static model is dynamic to build the classification matching push model. Combined with data set verification, it is concluded that the model is true and effective and can realize personalized recommendation.