Domingo of IICS presents paper on Breast Cancer prediction at int’l conference on Advance Information Science in Singapore

Institute of Information and Computing Sciences faculty member Asst. Prof. Mylene J. Domingo presented a paper titled “Fuzzy Decision Tree for Breast Cancer Prediction” during the 2019 International Conference on Advanced Information Science and System and the 2019 International Conference on Algorithms, Machine Learning and Signal Processing held from November 15 to 17, 2019 at the Nanyang Executive Centre, Nanyang Technological University, Singapore.   

In her paper, Domingo discussed the use of data mining to predict the stage of cancer using a hybrid of fuzzy logic and decision tree. Her research aims to help experts make decisions, instead of replacing them. The research provides recommendations, but the final decision still remains with the expert, according to Domingo, who once served as Secretary of IICS. 

Data mining is defined as the process of discovering patterns in large data sets involving methods at the intersection of machine learning, statistics, and database systems.

Decision tree is a classification algorithm that uses tree-like model of decisions. It handles both continuous and categorical attribute. Decision Tree algorithms uses IF-THEN expressions where the conditions are logically connected. 

Domingo explains that fuzzy logic has the capability to deal with problems with uncertainty similar to reasoning of experts. It is a logic system that deals with values between 0 and 1, which should be measured as degrees of truth while the decision tree that only deals with values as either 0 or 1.

For this research, Domingo used feature selection to determine the best attribute in the dataset from Surveillance Epidemiology and End Results (SEER). Although the data set consists of incidence from 1975 to 2016, the study limited its analysis from the year 2010 to 2016. Different cleaning and preprocessing of data were conducted. After thorough preprocessing of data, one target class and six attributes were selected. 

Performance comparison showed that the fuzzy decision tree achieved a higher accuracy of 99.96%, sensitivity of 99.26% and specificity of 99.98% than the decision tree classification technique. The simulation result showed a correctly classified instance of 165,124, which is equivalent to 99.97% and only 351 incorrect classified instances or 0.21%. Her research showed that a fuzzy decision tree is more robust than the traditional decision tree classifier for predicting the stage of breast cancer.  

Paper presenters were from China, Norway, Korea, Singapore, Turkey, India, Thailand, Myanmar, Saudi Arabia, Philippines, Indonesia, Malaysia, Ukraine, among others.

Like this article?

Share on Facebook
Share on Twitter