摘要
简要地介绍了在大规模数据库中挖掘关联规则的Apriori算法 ,给出了红外光谱数据库知识发现的空间表示方法 ,并根据红外光谱数据挖掘的特点改进了Apriori算法中支持度的计算与频繁集的确定过程 ,运用统计方法把挖掘结果形成可视的特征谱带 -化学基团规则式 ,通过具体的挖掘事例对挖掘结果进行分析与评价。挖掘出的规则式和波谱分析理论比较结果证明了挖掘结果的正确性 。
The Apriori algorithm of mining large database for associate rules is introduced simply. The space express method of knowledge discovery from infrared spectra database is proposed, and the computation method of support and procedure of frequent itemsets confirmed in Apriori algorithm are improved according to peculiarity of infrared spectra data mining, the analysis and evaluation of mining results are also presented by with of mining real infrared spectra data, the direction of next research work is disused in the end. The results of mining IR spectra are translated into visual rules of IR characteristic peak-substructure by statistic. The validity of mined rules is proved through mined rules compare with spectrum analysis theory, and it is effective and useful method to mine database of IR spectra by improved Apriori algorithm.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2001年第5期477-483,共7页
Computers and Applied Chemistry
基金
辽宁省自然科学基金 ( 9910 2 0 0 2 0 5 )
高校科研基金 ( 2 0 0 12 0 73)资助
