Learning Vector Quantization (LVQ) For Colorectal Cancer Identification Based on Microscopic Network Image

Colorectal is a type of malignant cancer that occurs on the surface of the large intestine (colon) and the lower part of the intestine to the anus (rectum) due to environmental influences and unhealthy lifestyles. The main function of the large intestine is to reabsorb water and to secrete mucus which serves to lubricate and help expel feces and gases. Colorectal cancer malignancy can attack anyone, from toddlers, teenagers, and adults. Identification of colorectal cancer is still using hispathological examination. Where this examination is a diagnostic act carried out by taking samples of cells or tissues for analysis in the laboratory. The examination is still done manually, namely using a microscope. In this way, a doctor who has the knowledge, thoroughness and accuracy is needed. This inspection takes time and effort. Therefore we need a way to help doctors identify colorectal cancer through microscopic images of colorectal cancer so that the identification results obtained are more efficient and have a better level of accuracy than manual identification. The identification process using input data includes the number of epohs as much as 100, the number of hidden layers of 3, the learning rate of 0.05. The test in this test uses colorectal tissue image testing data with the IMG18 filename. Produces Normal identification results with an accuracy of 60%, meaning that img18 is identified as normal colorectal.