A New Compound Method of Multi-dimensional Sequential Pattern Mining

Provided by: AICIT
Topic: Big Data
Format: PDF
Current multi-dimensional sequential pattern mining methods, which only mine multi-dimensional patterns from sparsely multi-dimensional database rapidly, are not very fast to mine multi-dimensional patterns in densely multi-dimensional database, especially when the dimensionality is higher. In this paper, the authors propose a new compound method Seq-Cmp for mining multi-dimensional sequential patterns, and define array-based structure H-arrays and tree-based structure H-forest. Seq-Cmp mines sequential patterns in dataset firstly, and then forms the corresponding projected multi-dimensional database for each sequential pattern, finally compounds to use H-arrays or H-forest to mine multidimensional patterns in sparsely or densely projected multi-dimensional database.

Find By Topic