Fig 1.
The exploited algorithm in this study.
Table 1.
Variables relating to the recurrence of breast cancer in a dataset.
Table 2.
Table 3.
Frequency distribution of demographic and clinical characteristics of patients.
Table 4.
Confusion matrix.
Table 5.
Frequency distribution of ER and PR receptors in terms of HER-2 receptors.
Table 6.
The test of different architectures of MLP neural network.
Table 7.
The result of the LVQ algorithm for the number of neurons.
Table 8.
A survey of different neural network architectures considering all input parameters in the recurrence of breast cancer disease.
Fig 2.
The workflow of the Raman spectral discrimination model of KPCA-SVM.
Fig 3.
Performance criteria of the seven classification methods.
Table 9.
Performance measure.
Table 10.
Actual and predicted recurrence in the studied patients.
Table 11.
Cancer recurrence matrix of the C5.0 algorithm.
Table 12.
Accuracy of the prediction model for each type of recurrence.
Table 13.
The total amount of disease-free survival in the first to fifth years after the initial treatment.
Table 14.
The most important feature with the highest accuracy in the diagnosis of breast cancer.