An Efficient Enhancement of Mining Top-K Association Rule

Data mining on huge databases has been a major issue in research area, due to the problem of analyzing large volumes of data using traditional OLAP tools only. This type of process implies much computational power, disk I/O and memory, which can be used only by parallel computers. So, depending on the selection of the parameters (the minimum support and minimum confidence), current algorithms can be slow and produce an extremely large number of results or produce very less results, omitting useful information. This is really a major problem because in practice users don’t have much resource for analyzing the results and have to discover a certain amount of results in a limited time. To address this problem, we propose a unique technique to mine top ranked data from a data set. The algorithm uses a new method for generating association rules. The algorithm has unique and best performance and feature of scalability, which is a beneficial alternative to classical Association rule mining algorithms when the user wants to control the number of rules generated. Keyword: Association Rule Mining, Data mining, Association rule learning, Top-k rules, Confidence