GPU-based multiple back propagation for big data problems
The big data era has become known for its abundance in rapidly generated data of varying formats and sizes. With this awareness, interest in data analytics and more specifically predictive analytics has received increased attention lately. However, the massive sample sizes and high dimensionality pe...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Published: |
International Center for Scientific Research and Studies
2016
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/66994/ http://eprints.utm.my/66994/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The big data era has become known for its abundance in rapidly generated data of varying formats and sizes. With this awareness, interest in data analytics and more specifically predictive analytics has received increased attention lately. However, the massive sample sizes and high dimensionality peculiar with these datasets has challenged the overall performance of one of the most important components of predictive analytics of our present time, Machine Learning. Given that dimensionality reduction has been heavily applied to the problems of high dimensionality, this work presents an improved scheme of GPU based Multiple Back Propagation (MBP) with feature selection for big high dimensional data problems. Elastic Net was used for automatic feature selection of high dimensional biomedical datasets before classification with GPU based MBP and experimental results show an improved performance over the previous scheme with MBP. |
|---|