©2019 Health Discovery Corporation

OUR TECHNOLOGY

SUPPORT VECTOR MACHINES

The Method to Success

Since their introduction, Support Vector Machines (SVMs) marked a breakthrough in the theory of learning systems.  Rooted in the Statistical Learning Theory, Support Vector Machines quickly gained attention from the pattern recognition community due to its theoretical and computational merits.

Statistical Learning Theory, the backbone of Support Vector Machines, provides a new framework for modeling learning algorithms, merges the fields of machine learning and statistics, and inspires algorithms that revolutionize the field of artificial intelligence. A new generation of learning algorithms - or equivalently of statistical methods - has recently been developed, based on this theory.

SVMs outperform even advanced statistical modeling methodologies such as neural networks.  Neural networks suffer from a limited ability to handle data, and can only analyze data from two or three dimensions.  Support Vector Machines, however, are able to process infinite amounts of data and analyze data to find separations and delineations within high dimensionality. 

Health Discovery Corporation's SVM Technology is commonly considered within the context of artificial intelligence. This is a branch of computer science concerned with giving computers the ability to perform functions normally associated with human intelligence, such as reasoning and optimization through experience. Machine learning is a type of artificial intelligence that enables the development of algorithms and techniques that allow computers to learn. Pattern recognition is machine learning with a wide spectrum of applications including medical diagnosis, bioinformatics, classifying DNA sequences, detecting credit card fraud, financial market analysis, and object recognition in digital images.

SVM-RFE

Exceptional Achievement

Created by Health Discovery Corporation's mathematicians, Support Vector Machine-Recursive Feature Elimination (SVM-RFE) is used, among other things, to discriminate relationships within clinical datasets. Within gene expression datasets created from micro-arrays of tumors, SVM-RFE distinguishes normal versus abnormal tissues. Utilizing SVM-RFE, Health Discovery Corporation is able to access specific genetic information that the most advanced bioinformatics techniques previously missed. For example, SVM-RFEs are able to filter irrelevant, tissue-specific genes from those related to malignancy.  SVM-RFE also identifies gene expression patterns related to the severity of a disease.  This data analysis technique provides the physician with patient-specific information and is an enhanced decision-making tool for pharmacogenic and toxicological profiling of the patient. This allows for the development of true personalized medicine. Our mathematicians believe Health Discovery Corporation's SVM-RFE analytic methods are effective for finding genes implicated in several cancers.

Additionally, SVM-RFE technology is able to rank order the analyzed genetic information to determine the most important gene in the molecular diagnostic test. The ability to rank order the genes also allows our mathematicians to identify IP free genes available for patent protection by Health Discovery Corporation.

The success of SVM-RFE is documented in numerous published academic papers worldwide.

Because SVM-RFE technology was identified by Health Discovery Corporation, we hold the only issued patents in the world for this technology. This was recently confirmed by the U.S. Patent and Trademark Office which ruled in favor of Health Discovery Corporation regarding the SVM-RFE patents as a result of an interference proceeding between Health Discovery Corporation and Intel Corporation. 

VALIDATION

Proven Technology

The Support Vector Machine is established as one of the leading approaches to pattern recognition and machine learning worldwide. In addition to their application in molecular diagnostics, SVMs are replacing neural networks and other methodologies in a variety of fields including engineering, information retrieval, and bioinformatics. They have received widespread notice for their ability to discover hidden relationships in datasets and have been exploited for numerous applications, including image analysis, text and image classification, handwriting and speech recognition, fingerprint identification, face recognition, fraud detection, financial securities prediction, protein structure prediction, medical diagnosis and prognosis, classification of microarray gene expression data, identifying consumer purchasing patterns, weather prediction, decryption, vehicle control monitoring, prediction of traffic variables in transportation systems, and many others.

Many commercial institutions have already incorporated SVMs into product and research applications. Industries utilizing SVM technology include large corporations within the software and computer fields, biomedical and pharmaceutical companies, automotive, financial, and retail industries. Educational and research institutions throughout the world have successfully applied SVMs to a wide array of applications including gene and protein expression analysis, medical image analysis and mass spectrometry.