The current work proposes a model for prediction of breast cancer using the classification approach in data mining. The proposed model is based on various parameters, including symptoms of breast cancer, gene mutation and other risk factors causing breast cancer. Mutations have been predicted in breast cancer causing genes with the help of alignment of normal and abnormal gene sequences; then predicting the class label of breast cancer (risky or safe) on the basis of IF-THEN rules, using Genetic Algorithm (GA). In this work, GA has used variable gene encoding mechanisms for chromosomes encoding, uniform population generations and selects two chromosomes by Roulette-Wheel selection technique for two-point crossover, which gives better solutions. The performance of the model is evaluated using the F score measure, Matthews Correlation Coefficient (MCC) and Receiver Operating Characteristic (ROC) by plotting points (Sensitivity V/s 1- Specificity).


Chromosome, Mutation, Cancer, Classification, Data Mining, Bioinformatics