GANDIVA
| Autor: | Vangipuram Radhakrishna, P. V. Kumar, Vinjamuri Janaki |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Tree (data structure)
Association mining General Computer Science Similarity (network science) Computer science business.industry 0202 electrical engineering electronic engineering information engineering 020207 software engineering 020201 artificial intelligence & image processing Pattern recognition 02 engineering and technology Artificial intelligence business |
| Zdroj: | International Journal of Information Technology and Web Engineering. 14:1-18 |
| ISSN: | 1554-1053 1554-1045 |
| DOI: | 10.4018/ijitwe.2019100101 |
| Popis: | In this research, the authors propose a novel tree structure called GANDIVA which computes true supports of all temporal itemsets by performing a tree-based scan and eliminating the database scan which is required for SPAMINE, G-SPAMINE, MASTER, and Z-SPAMINE approaches. The idea is to construct the tree called GANDIVA which determines support of all time-stamped temporal itemsets from the constructed tree. Another important advantage of the proposed approach is that it does not require the original database to be retained in the memory after a time profiled pattern tree (GANDIVA) is constructed from the original database. The significant advantage of GANDIVA over SPAMINE, G-SPAMINE, Z-SPAMINE, and MASTER is that GANDIVA requires zero database scans after the tree construction. GANDIVA is the pioneering research to propose a novel tree-based framework for seasonal temporal data mining. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|