A Matrix Model for Mining Frequent Patterns in Large Databases

This paper proposes a model for mining frequent patterns in large databases by implementing a matrix approach. The whole database is scanned only once and the data is compressed in the form of a matrix. The frequent patterns are then mined from this compressed database which brings efficiency in data mining, as the number of database scans is effectively less than two. The computation time is reduced as some of the patterns are mined simultaneously and searching is minimized. Appropriate mathematical operations are designed and performed on matrices to achieve this efficiency.