CAJ | 학술논문

Apriori算法是关联规则挖掘中最基本也是最重要的算法之一。但现有的Apriori算法存在对数据库的扫描次数过多，产生了大量的候选项集合，算法执行效率较低，运行时间显著增加等问题。文中针对现有Apriori扫描数据库过于频繁的问题，在Hadoop平台下对Apriori算法进行改进，提出一种改进算法H-Apriori算法。利用并行方法计算频繁项集，该算法将原始数据集按字母排序，减少频繁项集的计算开销，避免反复扫描数据库带来的时间上的消耗，从而提高算法的执行效率。通过与传统Apriori算法的执行时间相比较，实验结果表明，提出的改进算法H-Apriori明显减少了访问数据库的时间，有较高的执行效率。
Apriori산법시관련규칙알굴중최기본야시최중요적산법지일。단현유적Apriori산법존재대수거고적소묘차수과다，산생료대량적후선항집합，산법집행효솔교저，운행시간현저증가등문제。문중침대현유Apriori소묘수거고과우빈번적문제，재Hadoop평태하대Apriori산법진행개진，제출일충개진산법H-Apriori산법。이용병행방법계산빈번항집，해산법장원시수거집안자모배서，감소빈번항집적계산개소，피면반복소묘수거고대래적시간상적소모，종이제고산법적집행효솔。통과여전통Apriori산법적집행시간상비교，실험결과표명，제출적개진산법H-Apriori명현감소료방문수거고적시간，유교고적집행효솔。
Apriori algorithm is one of the most basic and important algorithms in association rules mining. But the existing Apriori algo-rithm scans database many times,generating a large number of candidate item sets,having a low execution efficiency,increasing running time significantly. Based on existing Apriori algorithm scanning database many times,put forward a new method,H-Apriori,which calcu-lates frequent items of association rules under Hadoop platform for improving efficiency of algorithm,reducing time accessing database, and sorts the original datasets,reducing frequent datasets’ computing cost,avoiding the time consumption by repeated scanning database, improving the efficiency of the algorithm. The experimental results show that the improved algorithm H-Apriori significantly reduces the amount of time to access the database with high execution efficiency.