Page 80 - Big Data Analytics for Intelligent Healthcare Management

4.6 RESULT AND ANALYSIS 73

Table 4.10 Result of Xception with LR, SVM, and K-NN on 40×

Feature Extractor Classifier Accuracy (%) Precision (%) Recall (%) F1-Score (%)
Xception LR 94.49 95.52 96.85 96.18
SVM 93.73 93.65 97.90 95.73
K-NN 91.48 92.08 94.94 93.49

100

98

96

94

92
90

88
Logistic regression Support vector K-NN

Accuracy Precision Recall F1score
FIG. 4.14
Performance of Xception with three different classifiers on 40 .

Interpretation: With Xception, Support Vector had the maximum recall but LR gave the best ac-
curacy, precision, and f1score (Table 4.10, Fig. 4.14).

4.6.3.2 Overall performance on 40×
Interpretation: The ResNet50 with both LR and Support Vector classifier had the maximum accuracy
but the Support Vector classifier had best recall value. On the other hand, with Inception ResNet V2,
Support Vector gave the highest precision (Fig. 4.15).

4.6.3.3 Test performance on 100×
Interpretation: With ResNet50, the Support Vector classifier gave the maximum recall value but K-NN
had the maximum precision while the LR gave the best accuracy and f1score (Table 4.11, Fig. 4.16).
Interpretation: With InceptionV3, the Support Vector classifier had the maximum recall value, but
the LR gave the best accuracy, precision, and f1score (Table 4.12, Fig. 4.17).