Intelligent Pre-diagnosis Method and Experimental Research Based on Patient's Personal Information

Chen Kai

《Journal of Clinical and Experimental Medicine》 2卷3期 (2018/09) Pp. 14-18

https://doi.org/10.29422/jocem.2018.03.003

摘要

Objective: Through the analysis of a large number of clinical data, and then explore the intrinsic relationship between symptoms, diagnosis and treatment records and diseases, it can construct a discriminating system for certain key diseases, which will be very beneficial to the doctor's diagnosis and treatment or the patient's self pre-diagnosis. Methods: To solve the problem of pre-diagnosis of the patient department, this problem was explored and tried using the rule-based method. Based on the results of the exploration, the classification algorithm of support vector machine and convolutional neural network was used to complete the multi-classification of the department. In the data preprocessing stage, it mainly introduces how to build a relational database between medical entities; how to standardize patient departments. For the purpose of effectively utilizing the patient's course record information to achieve accurate diagnosis and treatment, it is proved through experiments that what kind of diagnosis and treatment information is used is conducive to improving the correct rate of pre-diagnosis of which type of department. Conclusion: By comparing the experimental results with the evaluation results of the system, it can be seen that the intelligent pre-diagnosis method based on the patient's personal information has significantly improved performance compared with the pure rule and support vector machine.

關鍵字

clinical data ； intelligent diagnosis ； department pre-diagnosis ； convolutional neural network