Selection of suitable features using the GGA-PSO model for breast cancer diagnosis using the medical dataset of the state of Wisconsin, USA

artificial intelligence and machine learning tools, especially data mining, have come to the aid of the medical community to make the best decisions for the diagnosis and treatment of various diseases.Firstly, the combined method of two algorithms, GGA and PSO, was used, in the next step,the operation of combining algorithms with K-NN,random forest classifier algorithms for diagnosis and prediction and also as a selector of more useful and appropriate attributes are used. The obtained model is applied to three breast cancer datasets used,these datasets are Wisconsin Breast Cancer Database (WBC), (WDBC), and (WPBC).The values obtained in the feature selection method for WBC datasets, WDBC using the K-NN algorithm have shown high performance. Its value is 96.38% and 98.21%, respectively. Also, the amount of accuracy for the WPBC dataset using the random forest algorithm and the K-NN method. Its value is 92.11% and 89.47%, respectively. Finally,the obtained hybrid model was applied to WBC, WDBC and WPBC datasets, which selects the appropriate number of attributes to diagnose, prevent and predict whether the tumour is malignant or benign.Selecting the appropriate algorithms and combining them together can be useful in selecting the appropriate and efficient attribute. Accordingly, it can have a great impact on the criteria of accuracy,sensitivity, and accuracy in diagnosing and predicting whether breast cancer has been treated or recurred.

Selection of suitable features using the GGA-PSO model for breast cancer diagnosis using the medical dataset of the state of Wisconsin, USA Keywords:

Breast cancer , GGA algorithm , PSO algorithm , FEATUER SELECTION , K-NN algorithm.