Design And Implementation Of Health Monitoring Platform Based On Microblog Text

The rapid growth of social media has provided unprecedented opportunities to monitor the personal health reports of millions of people in real time.Health monitoring can play a positive role in disease research.Traditional survey-based monitoring methods have limited resources and have a certain lag.A large number of self-health report data on Weibo can be used as a good supplementary data in addition to the formal reports required by traditional monitoring methods,assisting relevant institutions in disease research,and helping to find seasonal,endemic,and epidemic diseases.A key step in achieving monitoring is to identify the tweets that express the author’s body with a particular symptom.This paper proposes a Weibo symptoms mentioned recognition model(WCCL＿MCNN)for Chinese short texts on Weibo.Text representation is constructed by joint representation model(WCCL)that combining text word level,word level,twisted centroid and extended feature based on centroid.Then,the text representation is input into the convolutional neural network to extract higher level semantic information and train a high-precision classifier to complete the task of mention recognition.Compared with the traditional model of text classification,text joint representation model can represent text from different levels,at the same time,based on the centroid extension features can improve the generalization ability of classifier.Finally,the convolutional neural network is used to extract the local features of higher level,and the text is classified through pooling,full connection layer and Softmax regression layer.Experiments show that the model in this paper can solve the problems of short text feature sparse and data imbalance to a certain extent,and has a good performance for the task of symptom mention recognition and classification of Weibo.This article also combines theory with practice,applies Weibo symptom mention recognition model,and develops a Weibo health monitoring system,which continuously and systematically collects Weibo users’ symptom reports,and finally displays the heat map of the national distribution of related symptoms,the specific location table of symptom reports,and the line chart of quantity trend,ect to users,so that they can clearly see the temporal and spatial distribution of relevant symptom reports on Weibo.Provide a large number of health report statistics for relevant researchers,assist in monitoring public health,disease research,and discover seasonal,endemic,and epidemic diseases.