期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2013
卷号：55
期号：1
出版社：Journal of Theoretical and Applied
摘要：Transaction database is used to find frequent itemsets from large datasets along with their associated timestamps. The existing frequent pattern mining algorithms such as Apriori, FP-growth, Partition, Pincer Search do not consider the timestamps associated with the transactions. In real time transactions, without timestamps it is difficult to identify the constantly changing behaviour of the frequent patterns in a transaction database. To overcome the above problem, EP-TP (an Efficient Periodic Transitional Patterns) mining method is introduced in this paper. Transitional patterns are used to discover transi-frequent patterns along with their time stamps in a transaction database. The above patterns include both positive and negative transitional patterns. A pattern is said to be a positive or negative transitional pattern if their frequencies dramatically increase or decrease respectively at some point of time in a transaction database. From the above, the business owners can able to analyze the trend of the pattern(s) over the period(s). This trend is used to find out the nature of the pattern and hence the business owner can able to make the necessary step to improve the performance (or analyze the reason) of the negative pattern. The above pattern is tested with Textile Dataset and its performance was compared with existing algorithm. It is also present the experimental study to verify the usefulness and effectiveness of transitional patterns.
关键词：Transaction database; Transitional pattern; Frequent pattern; Periodic pattern; Linear pattern
