A hybrid approach for decision making to detect breast cancer using data mining and autonomous agent based on human agent teamwork

Every year many women die of breast cancer and the saddest part is that most of them die due to late diagnosis. A number of studies have been carried out to explore the basic reasons of breast cancer. Unfortunately, most of them failed to detect the main causes and the disease itself at a primary stage. On the other hand, it has already been proved that early detection of cancer can give the patient a chance of longer survival. Therefore, early detection of breast cancer is very crucial to save a patients' life. To address this problem, we have introduced a hybrid model to identify breast cancer at primary stage. In this model, the first part includes data mining using decision tree algorithm. The second part includes an autonomous agent that takes decision based on predefined rules to detect breast cancer at the very beginning stage. These rules are deduced through a data mining tool (i.e. Weka).The autonomous agent has been developed using Java, which works in collaboration with human. In this research work, we mostly focused on creating an adjustable autonomous agent and setting its rules and behaviors effectively. The performance of the proposed hybrid model has been tested on breast cancer dataset collected from UCI (University of California, Irvine) machine learning repository. The study reveals that autonomous agent works better when it collaborates with human as a team member. In addition, this hybrid model can be used to assist medical practitioners to provide better treatment to the patients.