数据挖掘算法研究及其在处方分析系统的应用

发布时间：2018-01-01 21:17

本文关键词：数据挖掘算法研究及其在处方分析系统的应用　出处：《南京理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：中国医学是中华名族的瑰宝,中药作为中医的主要依托,在中医理论中发挥了巨大作用。千百年来中国人民的健康问题基本上是靠着看中医、吃中药解决的。随着中医院信息化的建设,医院数据库中存储了大量的中医处方。因此,利用兴起的数据挖掘技术,将传统的中医处方和现代的信息处理技术相结合,可以在没有先验知识的情况下,对药材之间的关系和处方中蕴含的规律进行挖掘,该研究有良好的应用前景。本文介绍了基于关联规则的经典Apriori算法,分析利用并行思想改进的计数分布算法(CD算法)。针对传统算法的不足,提出基于矩阵的并行关联规则挖掘算法(PM算法),解决传统串行算法速度的瓶颈。和CD算法相比,PM算法可以更快的生成频繁项集。本文分析了传统的粗糙集属性约简算法,通过对其建模分析,提出了基于模拟退火的属性约简算法,并将算法运用到条件属性较多的决策系统之中。针对模拟退火算法的随机性,优化算法产生新属性集的过程,加快了算法的收敛速度。基于上述研究成果,结合中医院的实际需求,设计了一个中药处方分析系统,实现对药材与药材、病证与药材之间关系的挖掘。通过实际处方数据比较不同算法的有效性,验证了该中药处方分析系统可以提供一定的用药指导,并且可以提高药房的工作效率,具有较高的实用价值。
[Abstract]:Chinese medicine is a treasure of the Chinese nation, Chinese medicine as the main support of traditional Chinese medicine, has played a huge role in Chinese medicine theory. For thousands of years, people's health problems Chinese basically is relying on Chinese medicine, eat Chinese traditional medicine to solve. With the construction of hospital information, hospital database stored in large quantities. So the prescription of Chinese medicine the rise, using data mining technology, the traditional Chinese medicine prescription and modern information processing technology, in the absence of a priori knowledge, to mine contains herbs and the relationship between the prescription rules, the research has a good application prospect. This paper introduces the classic Apriori algorithm based on association rules count distribution analysis using the theory of the improved parallel algorithm (CD algorithm). Aiming at the shortcomings of traditional algorithms, we propose parallel association rules mining algorithm based on matrix (PM algorithm), to solve the traditional serial The speed bottleneck algorithm. Compared with CD algorithm, PM algorithm can quickly generate frequent itemsets. This paper analyzes the traditional attribute reduction algorithms of rough set, through the analysis of the modeling, this paper presents a new attribute reduction algorithm based on simulated annealing algorithm, and the algorithm is applied to the decision-making system of condition attributes more. According to random simulation annealing algorithm, optimization algorithm to generate new attributes, accelerate the convergence speed of the algorithm. Based on the above research results, combined with the actual needs of the hospital, designed a prescription analysis system of medicinal materials and medicinal materials, exploring the relationship between the disease and medicine. The effectiveness of the actual data comparison of different prescription the algorithm, verified the prescription analysis system can provide a medication guide, and can improve the work efficiency of the pharmacy, it has a high practical value.

【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】