A Smart Data Pre-Processing Approach to Effective Management of Big Health Data in IoT Edge

Figures & data

Figure 1 Architecture used for anomaly detection and performance testing.

Figure 2 Simulation model.

Figure 3 Data source.

Table 1 Classification Performance Measures

Name	Formula
Error and Accuracy	(fp + fn)/N (tp + tn)/N = 1–error
TP-Rate	tp^/p
FP-Rate	fp/n
Precision	tp^/p
Recall	tp/p = tp-rate
Sensitivity	tp/p = tp-rate
Specificity	tn/n = 1 – fp-rate

Table 2 Random Cut Forest Classification Report

	Precision	Recall	F1-Score	Support
0	0.00	0.00	0.00	2782
1	0.07	1.00	0.14	218
Accuracy			0.07	3000
Macro Avg	0.04	0.50	0.07	3000
Weighted Avg	0.01	0.07	0.01	3000

Table 3 Logistic Regression Classification Report

	Precision	Recall	F1-Score	Support
0	0.52	0.98	0.68	1540
1	0.64	0.03	0.06	1460
Micro Avg	0.52	0.52	0.52	3000
Macro Avg	0.58	0.51	0.37	3000
Weighted Avg	0.57	0.52	0.37	3000

Table 4 Neural Networks Classification Report

	Accuracy Scores
Random Cut Forest	0.93
Logistic Regression	0.92
Neural Networks	0.0726
Naive Bayes	0.5193

Table 5 Naive Bayes Classification Report

	Precision	Recall	F1-Score	Support
0	0.94	0.99	0.96	2783
1	0.56	0.16	0.24	217
Micro Avg	0.93	0.93	0.93	3000
Macro Avg	0.75	0.57	0.60	3000
Weighted Avg	0.91	0.93	0.91	3000

Table 6 Accuracy Scores

	Precision	Recall	F1-Score	Support
0	0.94	0.99	0.96	2783
1	0.56	0.17	0.26	217
Micro Avg	0.93	0.93	0.93	3000
Macro Avg	0.75	0.58	0.61	3000
Weighted Avg	0.91	0.93	0.91	3000

Table 7 Comparison of Data Processing Speed

	Execution Time	CPU Time
Random Cut Forest	148 milliseconds	150 milliseconds
Logistic Regression	0.44 milliseconds	0.625 milliseconds
Neural Networks	41.800 milliseconds	67,000 milliseconds
Naive Bayes	1.67 milliseconds	2.19 milliseconds

Figure 4 Random cut forest roc curve.

Figure 5 Random cut forest auc score.

Figure 6 Logistic regression roc curve.

Figure 7 Logistic regression auc score.

Figure 8 Neural network roc curve.

Figure 9 Neural network auc score.

Figure 10 Naive Bayes roc curve.

Figure 11 Naive Bayes auc score.

Figure 12 Random cut forest confusion matrix.

Figure 13 Logistic regression confusion matrix.

Figure 14 Neural network confusion matrix.

Figure 15 Naive Bays confusion matrix.