Regression of Representative Keys for Classification: A Simple Learning Approach
El Sayed Mahmoud, David A. Calvert · ASME Press eBooks · 2009
Multidimensional data sets are common in many machine-learning applications such as in the detection of disease outbreaks. Learning complexity is related to the high dimensionality of the data. Researchers focus on hybrid solutions to deal with this type of complexity. This work chooses to simplify the data sets and the learning algorithm to address this problem. This work proposes a simple learning approach to detect disease outbreaks based on Emergency Department and Telehealth data. The proposed framework is not limited to a specific application. The key innovations are to simplify data representation by dimensionality reduction, and to simplify the learned model to be a two-column table. The main objectives are: (1) to develop a simple supervised learning approach; (2) to detect disease outbreaks using the approach; and (3) to compare results of detecting disease outbreaks with other approaches that have been used for the same problem.