INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
Improvised Feature Subset Selection Algorithm (FAST) for High Dimensional Data
Authors Name:
Priyanka Mate
, G.P. Chakote
Unique Id:
IJSDR1805041
Published In:
Volume 3 Issue 5, May-2018
Abstract:
In choosing a feature, we are concerned about finding those features that produce results similar to the original set of features. We take efficiency and effectiveness into consideration while evaluating the feature selection algorithm. Efficiency in dealing with the time needed to find a subset of features and performance with the quality of a subset of features. These criteria have introduced the FAST (FAST) Advanced Feature Selection Grouping and have been evaluated and used in this document. Reducing the size of data is one of FAST's most important features. First, we use group-graphing theories to segment properties. We then create a subset of features by selecting the most representative features and most relevant to the target classes. Because of the features in the groups are quite independent. FAST's grouping strategy is highly likely to provide a subset of useful and independent features. Specifies a subset of the most useful features that produce a compatible result because all feature sets are involved in the feature selection. The attribute selection algorithm can be evaluated from the performance point of view and its effectiveness. Performance is related to the quality of a subset of features, performance relative to the time it takes to find a subset of features.
"Improvised Feature Subset Selection Algorithm (FAST) for High Dimensional Data", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.3, Issue 5, page no.295 - 299, May-2018, Available :http://www.ijsdr.org/papers/IJSDR1805041.pdf
Downloads:
000336256
Publication Details:
Published Paper ID: IJSDR1805041
Registration ID:180227
Published In: Volume 3 Issue 5, May-2018
DOI (Digital Object Identifier):
Page No: 295 - 299
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn