CAJ | 학술논문

文章阐述了FP—growth算法的特征，并针对该算法的不足，提出了基于分解事务矩阵的FP—growth改进算法。该算法利用压缩的布尔矩阵对事务信息进行存储，仅扫描数据库一次，当事务数和支持度阙值发生变化，可不必重新扫描数据库，特别是当数据集很大时，能有效提高时间和空间的挖掘效率。
문장천술료FP—growth산법적특정，병침대해산법적불족，제출료기우분해사무구진적FP—growth개진산법。해산법이용압축적포이구진대사무신식진행존저，부소묘수거고일차，당사무수화지지도궐치발생변화，가불필중신소묘수거고，특별시당수거집흔대시，능유효제고시간화공간적알굴효솔。
This paper describes the features of FP - growth, aiming at the deficiency ot the algorithm, puts forward improved FP- growth algorithm based on transaction matrix decomposition. The algorithm uses a compression of the Boolean matrix on the transaction information is stored only once, scanning the database, when the number of transactions and support value changed, but does not need to scan the database, especially when the data set is very large, can effectively improve the mining efficiency of time and space