Table 3.

A comparison of the discussed anomaly detection schemes.

Scheme	Technique	Prior Knowledge	Complexity	Practical Consideration	Accuracy	Data Prediction	Anomaly	Drawback
Xie et al. [34]	Multivariate	Yes	Low	No	High	No	DOS	Dimensionality/PK
Magan-Carrion et al. [35]	Multivariate	Yes	Low	Yes	High	Yes	Data Loss/Modification	Affected by Routing/PK
Magan-Carrion et al. [36]	Multivariate	Yes	Low	No	High	Yes	Tampered Data	Traffic Imbalance/PK
Xie et al. [41]	kNN	No	Moderate	No	High	No	Generic	No Regression
Liu et al. [42]	kNN	Yes	High	No	High	No	Generic	Dimensionality
Zhu et al. [43]	kNN	No	High	No	High	No	Misbehaving Nodes	Complexity
Martins et al. [45]	SVM	No	High	No	High	No	Generic	Complexity
Salem et al. [46]	SVM	Yes	High	No	High	Yes	Data Integrity	Complexity/PK
Shilton et al. [47]	SVM	Yes	High	No	High	No	Generic	Complexity/PK
Cannady [50]	ANN	No	High	No	N/A	No	DOS	Complexity
Bosman et al. [51]	ANN	No	Moderate	Yes	High	No	Generic	Detection Bias
Yusuf et al. [52]	ANN	Yes	High	Yes	High	Yes	Data Integrity	Complexity/PK
Radhika et al. [53]	GA	Yes	High	No	Average	No	Misbehaving Nodes	Complexity/PK
Bankovic et al. [54]	GA	No	High	Yes	High	No	Misbehaving Nodes	Complexity
Rizwan et al. [55]	GA	Yes	High	No	High	No	Generic	Complexity/PK
Maleh et al. [56]	Hybrid	Yes	Moderate	Yes	High	No	DOS	PK
Ma et al. [57]	Hybrid	Yes	High	No	High	No	Generic	Complexity/PK