SAVANNAH, Ga.--(BUSINESS WIRE)--Health Discovery Corporation (HDC) (OTCBB:HDVY) today announced that the U.S. Patent and Trademark Office (USPTO) has issued a notice of allowance for HDC’s patent application covering a number of different feature selection techniques that can be used either in conjunction with or as an alternative to HDC’s SVM patents using recursive feature elimination (RFE-SVM). These include minimizing the number of non-zero parameters of the system (or “zero-norm minimization”), evaluation of cost function to identify a subset of features compatible with constraints imposed by the learning set, unbalanced correlation score and transductive feature selection. It also describes a method that allows direct processing of multi-class problems instead of the more common practice of decomposing the problems into a multitude of two-class scenarios, thus providing for feature selection in the realm of multi-class and categorical regression problems. Examples are described for identifying the most relevant genes for classification of colon cancer and prostate cancer.
