The Implementation and Optimization of Public Emotion Network Communication Model by Deep Learning

To study the application of deep learning algorithms in the direction of public sentiment analysis, this study takes Weibo, the representative of the public sentiment communication platform, as an example and proposes a text sentiment analysis method based on sentiment dictionary and machine learning. To make up for the shortcomings of the convolutional neural network in the fully connected classification layer, which has a weak classification ability of nonlinearly separable data, the inseparable nonlinear problem combined with the support vector machine is transformed into a separable linear problem. The emotion analysis model of microblog short text based on convolutional neural network and support vector machine is established and combined with deep learning and artificial intelligence technology. In the Weibo database, actual data are extracted for model validation.