Feature and Instance selection via cooperative PSO
Advances in data collection and storage capabilities during the past decades have led to an information overload in most application domains. The huge amount of data the real-world applications has necessitated the use of a reduction mechanism. The reduction method contains two main techniques: feat...
Saved in:
| Main Author: | |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | en |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/11359/1/SSSA2012_2.pdf http://eprints.utem.edu.my/id/eprint/11359/ http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6083986&tag=1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Advances in data collection and storage capabilities during the past decades have led to an information overload in most application domains. The huge amount of data the real-world applications has necessitated the use of a reduction mechanism. The reduction method contains two main techniques: feature selection and instance selection, which are usually applied individually. Although, some work has been done to implement the feature and instance selection simultaneously, this work has focused on mainly the classification problem. This paper proposes the integration of feature selection and instance selection for solving the regression problem by using the fuzzy modeling approach. The selection of features and instances is based on the cooperative particle swarm optimization technique, which aims to limit the effect of the curse of dimensionality that occurs when dealing with the high dimensionality of the search space. The proposed method is applied to three real-world datasets from the machine learning repository. The algorithm's performance is illustrated by the corresponding plots of the prediction error for the different amounts of data being selected. |
|---|