FSKY-Miner: Fast Mining of Skyline Patterns.

Frequent pattern mining is one of the major tasks to ascertain the frequency of itemsets in a transaction database. Utility-driven pattern discovery takes into account additional factors like interestingness or unit profit. However, we may consider more factors. For instance, while choosing hotels, they may carefully examine both price and distance. In recent years, skyline pattern mining, which focuses on frequency and utility aspects, has garnered substantial attention, and several algorithms have been developed, enabling people to make better decisions when considering multiple factors. The issue of mining skyline patterns from different perspectives is explored in further detail in this work. We provided an efficient algorithm called FSKY-Miner with an effective initialization strategy, two pruning strategies involving subtree utility and local utility, and projected database as well as transaction merging technologies, which collectively aim to efficiently prune unpromising itemsets for the purpose of avoiding generating large numbers of candidates. To demonstrate the comparison results of FSKY-Miner with cutting-edge algorithms and prove the impact of different strategies on FSKY-Miner, numerous experiments were conducted. FSKY-Miner performs better than other algorithms in terms of runtime, memory usage, search space, and scalability.