Cite this article:
Verma and Vyas (2005). Concept of Decision Tree Classifier for Temporal data. Journal of Ravishankar University (Part-B: Science), 18(1), pp.07-22.
Journal of Ravishankar University Vol.18 No. B (Science) 2005 pp 07-22 ISSN 0970-5910
Concept of Decision Tree Classifier for Temporal data
Keshrj
Verma and O.P. Vyas
School of Studies
in Computer Science Pt. Ravishankar Shukla University, Raipur
Chhattisgarh . 492010
seerna2000•@yahoo.com,opvyas@rediffmail.com
MS Received : 03/08/04 Revised : 28/04/05 Accepted : 03/06/05
Abstract : Classification is an
important problem in data mining Given a
database of records, each with a class label, a
classifier generates a concise and meaningful
description for each class that can· be used to classify future
records whose classes are unknown.
A
number of populat classifier exists to construct decision tree and to generate class model. This Classifier first builds decision tree and then prunes the subtrees from the decision tree in a subsequent pruning phase
to improve accuracy and prevent overiitting. After the prunining
the number of nodes is if very large for large data
set, it takes more time to classify the data. Hence, the time component is one
the most important factors
of real world dataset. Each data has the specific time
period in which the fact is true. It is called valid time. The present proposes to extend and improve the existing SPRINT algorithm by incorporating the temporal aspects. This classifier has the several advantages over the existing onces. It reduces the size of dataset by partitioning
the data into different intervals. As a resultant, it takes less space. It acts as a faster classifier, which decreases time complexity. A chronological development ofclassifier
is displayed at the end.
Keyword : Classification rule,
Supervised learning, Temporal classification.
Decision Tree Classification no : H.2.8. (ACM Classification scheme)
NOTE: Full version
of this manuscript is available in PDF.