13-09-2017, 09:40 AM
Frequent patterns are very important in the discovery of knowledge and in the process of data mining, such as mining association rules, correlations. FP-tree is a very versatile data structure used for the extraction of frequent patterns in the discovery of knowledge and the process of data mining. FP-tree is a compact representation of the transaction database containing frequency information of all relevant relevant patterns (FP) of the database. All existing incremental frequent pattern mineralization algorithms, such as AFPIM, CATS, CanTree, CP-tree and SPO-tree, perform incremental mining by processing a transaction from the incremental part of the database at a time and updating it to the FP-tree of the initial (original) database. Here, in this article, we propose a new method that takes advantage of the incremental transaction database tree FP representation of incremental mining. We propose a batch incremental processing algorithm BIT_FPGrowth that restructures and merges two FP trees of small consecutive duration to obtain an FP tree of the FP-Growth algorithm. Our BIT_FPGrowth uses the FP tree as a preprocessed data repository for transactions (that is, sets of elements), unlike other sequential incremental algorithms that read database transactions. The BIT_FPGrowth algorithm takes less time to build FP-tree.
Frequent model mining has been the subject of numerous studies, including incremental updating. Many existing incremental mining algorithms are based on Apriori, which are not easily adopted for extraction of frequent patterns based on FP trees. In this paper we propose a new tree structure, called CanTree, which captures the contents of the transaction database and orders the tree nodes according to a canonical order. By leveraging its good properties, CanTree can be easily maintained when database transactions are inserted, deleted and / or modified. For example, CanTree does not require tuning, merging, and / or splitting tree nodes during maintenance. You do not need to re-browse the entire updated database or rebuild a new tree for incremental updating.
Frequent model mining has been the subject of numerous studies, including incremental updating. Many existing incremental mining algorithms are based on Apriori, which are not easily adopted for extraction of frequent patterns based on FP trees. In this paper we propose a new tree structure, called CanTree, which captures the contents of the transaction database and orders the tree nodes according to a canonical order. By leveraging its good properties, CanTree can be easily maintained when database transactions are inserted, deleted and / or modified. For example, CanTree does not require tuning, merging, and / or splitting tree nodes during maintenance. You do not need to re-browse the entire updated database or rebuild a new tree for incremental updating.