Research On Concept Extracition Technology Based On Collaborative Learning

With the rapid development of information science and computer technology,the process of information technology has also been rapidly advanced,a variety of related information systems used in mainstream institutions and management enterprises,especially medical information system in the use of electronic medical records.There are a lot of conceptual information in these electronic medical records,especially the secondary use of electronic medical records has been the concern of the majority of experts and scholars.The full use of electronic medical records related to the patient and has been marked with the concept of information,will promote the concept of medical technology to identify the development of technology,for this academia began to carry out relevant research,especially for medical electronic medical records of the secondary development and utilization The In view of the Chinese electronic medical records in the concept of the status of the identification of a total of the following studies:(1)The development of labeling norms and the construction of medical concepts marked corpus.According to the current relevant norms in China,combined with the ambiguity and complexity of the Chinese medical concept information itself,the concept of the concept of Chinese electronic medical records is finally formulated.In order to improve the efficiency of the document,we constructed a Chinese electronic medical record information labeling system.The system according to the specification of the electronic medical records in the location information and symptom information were marked,to take the system sampling and other methods have been marked with the quality of the document to ensure that the accuracy and reliability of the notes.(2)Research on medical concept information extraction based on single classifier.At present in the medical concept of information extraction method used,usually based on the machine learning method to complete the task,these methods to use the main monitoring method.In the study of medical concept information extraction,the hidden Markov model,the maximum entropy Markov model and the conditional random field model are used to complete the corresponding concept extraction task.Experiments show that the conditional stochastic field model is optimal in these three classifiers,but the three kinds of single class experiments do not consider the use of multi-classifier for information extraction tasks,and the use of multiple subsets for iterative processing.(3)Research on conceptual extraction of collaborative learning based on multi-classifier.Due to the limitations of the single classifier algorithm,a new algorithm is proposed to extract the information algorithm based on multiple classifiers.The new algorithm not only uses multiple classifier models,but also divides the data sets into multiple subsets.The subset is subjected to iterative training under different classifiers.Compared with the concept of three single classifier concept extraction experiments,we can see that the conceptual extraction algorithm is not only able to extract concept information efficiently and accurately,but also can identify new concept information.Therefore,we can see that the multi-classifier based collaborative learning concept extraction technology is more robust and applicable.