A robust hybrid machine learning algorithm for detecting breast cancer with reduced features

Abstract

Cancer is one of the most deadly disease which leads to death in human. It is newlinenothing but uncontrollable growth of cells in any particular part of the body. There are newline200 types of cancer and all the symptoms will not be cancer. It can also be a noncancerous newlinehealth conditions. There are many advanced automatic diagnosis systems newlinefor diagnosing different kinds of cancer. The breast cancer affects women mostly newlineafter the age of 50+ years or after menopause stage. The general symptoms of breast newlinecancer are Lump or thickened area in breast tissue, scaly patches in the skin etc. But newlineall lumps are not malignant tumors. This can be identified by using various diagnostic newlinemethodologies such as by using mammographic images, Fine Needle Aspiration newline(FNA), Scinitimammography images etc. Many diagnostic systems are available, but newlinemany researchers are still performing research on diagnosing breast cancer in a most newlineeffective way with a less computational cost. Many applications are developed as a newlinediagnostic system and recently researchers are concentrating on machine learning newlinewhich consists of many algorithms which can be effectively applied for diagnosing the newlinedisease easily. newlineMachine learning is one of the subfield of Computer Science. It was developed newlinefrom the study of pattern recognition. It is also a part of statistics and also it is closely newlinerelated to Linear algebra, Mathematic optimization, Matrix theory etc. Machine newlinelearning is also related to data mining and it is used in various fields such as newlinemarketing, online advertising, speech and hand written recognition, internet fraud newlinedetection, medical field etc. There are many applications designed using machine newlinelearning and in medical field many automatic systems are designed for detecting newlinevarious diseases. The doctors prefer the diagnostic systems which are user friendly newlineand also the system with a good accuracy.