Design And Implementation Of Intelligent Medical System Based On Text Classification

In recent years,artificial intelligence technology is becoming more and more mature.Machine learning,neural network and other technologies are increasingly applied in various fields,and have made great achievements.The medical sector belongs to knowledgeintensive field.It will produce massive data every day,so it can meet the demand for data volume in artificial intelligence.However,at present,there are few successful cases of the combination of artificial intelligence technology and the medical industry.Therefore,how to apply artificial intelligence technology to the medical industry,this topic has great academic value and application value.In view of this situation,this paper has done the following research:(1)Design a text classification model based on multi-head self-attention and capsule network(Capsule Net).In this model,multi-head self-attention mechanism is combined with multiple capsule networks to extract text features from different dimensions.Then,bidirectional long short-term memory(Bi-LSTM)is connected to classify text.(2)Design a text classification model based on the bidirectional encoder representations from transformers(BERT).The model first uses the BERT model to process the word vector into the capsule network and the bidirectional gate unit(Bi-GRU),and then fuse the output of multiple models to obtain the result of the text classification.(3)Develop an intelligent medical system that combines the text classification model with the online medical system.The system adopts the micro-service architecture mode,and includes user management module,medical module,forum module and text classification model module.In the text classification model module,pre-trained text classification models can be imported.Doctors or patients can assist in solving relevant problems through the models.Besides,the system collects data related to the newly generated model in a fixed cycle,and then the new data is labeled by the domain experts.After the data is labeled,it can be used for model retraining to ensure timeliness of the model.