As the fourth largest disease in the world,depression seriously affects people's physiological and mental health.To apply natural language processing techniques to automatically detect depressed people,we propose a de-pression recognition model combining text summarization and emotion perception.First,we use the text summariza-tion method to extract the global semantic features.Then we apply vocabulary enhancement methods to extract sen-tence-level emotional representation.Finally,we use deep neutral network to capture the emotion features.The re-sults show our model achieves up to 75.61％positive F1.
depression recognitionnatural language processingtext summarizationemotion perception
维普
万方数据
