FP-Tree算法规则挖掘的研究与应用

Research and application of FP-Tree algorithm rule mining

对FP-Tree算法的规则挖掘以及阈值设定与规则获取的关系进行了研究.选取高校医疗系统中存储的大学生体检数据,并对这些原始数据进行过滤、转换等加工处理,得到便于进行规则挖掘的事务数据库.将事务数据库中的数据用FP-Tree算法进行处理,得到数据之间的关联关系,从而对应获取大学生群体中常见慢性病之间的关联关系.在FP-Tree算法应用过程中设定相关参数的不同阈值,并反复实验调整最小支持度阈值和最小置信度阈值以满足医学标准.所获得的关联关系可以在患某种慢性病的早期就敦促大学生改掉不良嗜好、养成良好的生活习惯,降低严重慢性疾病发生的概率.

The rule mining of FP-Tree algorithm and the relationship between threshold setting and rule acquisition are studied.The research chooses the college students physical examination data stored in the medical system,filters and transforms the original data,and obtains the transaction database which is convenient for rule mining.The data in the transaction database are processed by FP-Tree algorithm,and the association relationship between the data is obtained,which corresponds to the relationship between common chronic diseases among college students.Different thresholds of relevant parameters are set in the application process of the FP-Tree algorithm,and the minimum support threshold and the minimum confidence threshold are repeatedly adjusted to meet medical standards.The association obtained can urge college students to change bad habits and develop good habits in the early stage of a chronic disease,so as to reduce the probability of more serious chronic diseases.