Pendiskretan data set kasar menggunakan ta’akulan boolean
Data discretization of rough set towards real attribute values is one of the important aspect in the data mining concepts, particularly the ones which involved classification problems. Emphirical results showed that the quality of the classification depends on the discretization algorithm used in...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Penerbit UKM
2004
|
| Online Access: | http://journalarticle.ukm.my/2159/ http://journalarticle.ukm.my/2159/ http://journalarticle.ukm.my/2159/1/vol1_abstract2.pdf |
| Summary: | Data discretization of rough set towards real attribute values is one of the important
aspect in the data mining concepts, particularly the ones which involved classification
problems. Emphirical results showed that the quality of the classification depends on the
discretization algorithm used in the input data pre-processing phase. In general,
discretization is a process of search-ing for partition of attribute domains into intervals
and unifying the values over each interval. Discretization process involves searching of
cuts which determine the intervals acquired. All values which lie within each interval are
mapped to the same values, in effect converting numerical attributes that can be treated
as being symbolic. The search for cuts is performed on the internal integer representation
of the input decision system. This paper describes the role of rough set with Boolean
reasoning discretization process in converting the real values of printed mathematical
symbol that leads to the better recognition rates using neural network |
|---|