Research for the Colon Cancer Based on the EMD and LS-SVM

On the account of the colon cancer genetic data signals are non-stationary and difficult to obtain a large number of cancer samples in the actual conditions, this text proposes a diagnostic research method of colon cancer based on EMD (empirical mode decomposition) combining with LS-SVM. Adopting the method of EMD decomposes the genetic signals of normal people and colon cancer patients. Figure out each IMF´s characteristic vector and regard them as the testing and training samples of LS-SVM. The testing results by classifying the colon cancer patients and normal people respectively show that the research approach of colon cancer based on EMD and LS-SVM has better classification results, which can achieve the goal of classifying in case of less samples. Those diagnosable and researching approaches propose a new method for the recognition of cancer.