MACFP: Maximal Approximate Consecutive Frequent Pattern Mining under Edit Distance.

Area of Science:

• Bioinformatics
• Computational Biology
• Data Mining

Background:

• Frequent pattern mining is crucial for biological sequence analysis, time series, and network logs.
• Existing methods often overlook insertions/deletions (indels), limiting approximate string pattern discovery.
• High computational complexity hinders approximate pattern mining under edit distance, especially for large DNA sequences.

Purpose of the Study:

• To address the challenge of approximate consecutive frequent pattern mining under edit distance.
• To develop an efficient algorithm for identifying substring patterns with indels in long sequences.
• To reduce the computational burden of Maximal Approximate Consecutive Frequent Pattern Mining (MACFP).

Main Methods:

• Formulation of the Maximal Approximate Consecutive Frequent Pattern Mining (MACFP) problem.
• Proposal of a novel algorithm with linear time complexity for support threshold checking.
• Integration of indexing and searching techniques for efficient pattern discovery.

Main Results:

• A significant reduction in computational complexity for approximate consecutive frequent pattern mining.
• Demonstrated effectiveness and efficiency through comprehensive experiments on sequence pattern analysis.
• Successful application in a cancer genomics study, showcasing practical utility.

Conclusions:

• The developed algorithm provides an efficient solution for approximate consecutive frequent pattern mining under edit distance.
• The approach overcomes limitations of existing methods by effectively handling indels.
• The algorithm shows promise for applications in bioinformatics and genomics, particularly with large datasets.