Industrial and Systems Engineering professor's article one of top 25 cited articles in European Journal of Operational Research

Alper Murat's, associate professor of industrial and systems engineering, article "A discrete particle swarm optimization method for feature selection in binary classification problems," co-authored with Alper Unler, is recognized as the one of the Top 25 Most Cited Articles in the European Journal of Operational Research. The article was cited over 100 times and ranks 17th out of 3,750 articles published in the European Journal of Operational Research since 2010.

This paper develops innovative models and methods that enable making highly accurate predictions from vast amounts of data. One of the most important challenges affecting business world and healthcare systems is the ability utilize immense amounts of data collected from various sources.

"Unfortunately, the majority of the existing methods are not capable of dealing with big data and hence require some selection of which features and attributes to utilize in making predictions. For example, it is important to consider patient's zip code in predicting a treatment alternative's outcome," said Murat.

The selection of which features and attributes to include is a very complex problem, especially with thousands of features and attributes, and critical to the performance of predictive analytics. Murat's paper introduces a dynamic intelligent selection concept in efficiently selecting the key and important features such that the prediction models built with these features are highly accurate.

To learn more about Murat's research visit https://engineering.wayne.edu/profile/ekrem.murat/.