A BELIEF RULE BASED EXPERT SYSTEM TO ASSESS COLORECTAL CANCER UNDER UNCERTAINTY

Colorectal cancer (or bowel cancer) is the fatal disease for both men and women in worldwide. The risk of getting colorectal cancer increases with age, most commonly after age 50. The highest rates are seen in male, whereas the rates of women are low. In Bangladesh, 12% adult male and 3% female are affected by colorectal cancer. Colorectal cancer is diagnosed from its various signs and symptoms by physicians. Maximum patients are unable to express their signs and symptoms to the doctors clearly. For this reason it became difficult to detect the colorectal cancer accurately. From their opinions (patients) various types of uncertainty (Unknowable information) issues arises. This diagnosis process cannot provide 100 % result. In this paper, an expert system using evidential reasoning approach (RIMER) is used to detect the presence of colorectal cancer. Using this expert system, we can handle different kinds of uncertainty. This research focused on the results of three systems namely Belief Rule Based Expert System, fuzzy logic based system and human experts. The web based expert system is used instead of desktop expert system which is user friendly. The data analyzed has been accomplished by using SPSS 22 and MATLAB. The output of the BRBES provides more reliable result than fuzzy logic expert system as well as human experts.