关键词: 临床路径, 优化, 遗传算法, Word2vec, 变异反馈

Abstract: To handle the problem of clinical pathway optimization,a method based on genetic algorithm was developed,in which a set of medical orders was represented by one chromosome.For low efficiency of the unadjusted genetic mining algorithm,the common optimization methods of genetic algorithm were investigated and applied to the optimization of clinical pathway.After that,their effects were compared.For further optimization of the genetic algorithm,an optimization method of mutation operator based on word2vec was developed.The experimental result showed that the genetic mining algorithm was feasible in the problem of clinical pathway optimization,and the effect of mining algorithm reached a maximum by using random sampling,fitnessscaling,elitist strategy,uniform crossover,adaptive mutation and the mutation method based on word2vec.The efficiency of optimized genetic mining algorithm was improved by 50%～55% with nonnumerical coding and about 77% with binary coding.

Key words: clinical pathway, optimization, genetic algorithms, word2vec, variations