Utilization of Stockwell Transform and Random Forest Algorithm for Efficient Detection and Classification of Power Quality Disturbances

Table 12

Comparing the effectiveness of the proposed method in view of other recently published articles.

Type of classification	Number of classes	Number of features	Sampling frequency	Overall classification accuracy (%) with noise
SSD + HD [26]	7	5	20 kHz	95.4
FFT + ELM [27]	12	7	6.4 kHz	95.38
HHT + WBELM [25]	16	4	3.2 kHz	95.6
EWT + SVM [21]	16	6	6.4 kHz	95.56
DWT + ML [28]	6	10	25 kHz	95.83
RF + DWT [29]	21	9	3.2 kHz	96.48
WT + PCA + SVM [30]	5	5	—	96.2
Proposed ST + RF	17	9	3.2 kHz	99.01

Note. SSD: sparse signal decomposition; HD: hybrid dictionary; HHT: Hilbert–Huang transform; WBELM: weighted bidirectional-based extreme learning machine; ELM: extreme learning machine; PCA: principal component analysis; SVM: support vector machine; FFT: fast Fourier transform; ST: S-transform; RF: random forest.