Molding Machine Learning for Business
Among machine learning techniques, classification
techniques are useful for various business applications, but classification
algorithms perform poorly with imbalanced data.
In their study appearing in a 2022 issue of Information Systems Frontiers, Turner College assistant professor
of management information systems Yoon
Lee and Chris Bang of Auburn University – Montgomery propose a
classification technique with improved binary classification performance on
both the minority and majority classes of imbalanced structured data. Their technique involves three steps,
including creation of a balanced training set via under-sampling and converting
it into an image depicting a line graph.
The third and final step in the process includes training a
Convolutional Neural Network, which is a Deep Learning algorithm
that can absorb an input image, assign importance to various objects in the
image, and differentiate one from the other, using the images. As Lee explained to Turner Business, “For the experiments, we applied the proposed
framework to six datasets from the University of California – Irvine
Repository. The proposed model achieved
the best receiver operating characteristic curve on all six datasets, and
Balanced Accuracy on five of the six datasets.”
Thus, Lee’s experimentation demonstrates that the combination of
under-sampling and Convolutional Neural Networks is a viable approach for
imbalanced structure data classification.
Comments
Post a Comment