Skip to main content
Fig. 8 | BMC Bioinformatics

Fig. 8

From: Implementation of ensemble machine learning algorithms on exome datasets for predicting early diagnosis of cancers

Fig. 8

Confusion matrix heatmap of neural network with undersampling. The primary diagonal elements from this graph shows the true correct positives and the rest are the false classification. Higher number of primary diagonals from the matrix shows that the classifier has achieved a good accuracy but performance was worse compared to SMOTE oversampling. 0–4 represents the five cancer classes. 0: High-grade serous ovarian cancer; 1: Human diffuse-type gastric cancer; 2: Intrahepatic cholangiocarcinoma; 3: Non BRCA1/BRCA2 familial breast cancer; 4: Pancreatic adenocarcinoma. The light to dark color coding indicates the probabilities of true and false positives

Back to article page