Table 8.

Comparison of the proposed model MSDA-DM with the recent studies on disaster datasets. The parameters used for comparison are Accuracy (Acc.), Precision (P), Recall (R), and F1-score (all in %)

Ref.	Approach	Source Disaster(s)	Target Disaster	Acc.	P	R	F1-Score
[10]	Fine-tuned BERT	The combined dataset of Hurricane Harvey, Hurricane Maria and		75.37	-	-	-
	CNN GRU	Hurricane Irma		67.63	-	-	-
	RCNN			71.77	-	-	-
	BiGRU			67.88	-	-	-
	CNN LSTM			58.09	-	-	-
	KMax CNN			66.50	-	-	-
	DPCNN			63.74	-	-	-
	BiLSTM			71.02	-	-	-
[49]	Binary classification of text and image tweets based on fusion	Hurricane Harvey		77.70	-	-	77.60
		Hurricane Maria		72.96	-	-	72.84
		Hurricane Irma		73.82	-	-	73.55
		Mexico Earthquake		74.29	-	-	74.25
		California Wildfires		65.00	-	-	64.00
		Iraq-Iran Earthquake		68.18	-	-	67.00
[35]	ANN and CNN	Hurricane Harvey		75.90	76.00	76.00	76.00
[2]	Semi-Supervised DA with Graph Embeddings and adversarial training to classify the text tweets into two classes: relevant and non-relevant	Nepal Earthquake	Queensland Earthquake	-	67.48	65.90	65.92
		Queensland Earthquake	Nepal Earthquake	-	58.63	59.00	59.05
[3]	Semi-Supervised inductive Learning to classify the text tweets into two classes relevant and irrelevant	Nepal Earthquake		-	-	-	52.32
		Queensland Earthquake		-	-	-	75.08
[25]	Semi-supervised DA for the binary classification of image tweets	Hurricane Harvey	Hurricane Irma	77.90	77.88	77.97	77.90
		Hurricane Harvey	California Wildfires	79.87	79.11	77.27	77.90
[33]	Unsupervised Domain Adversarial Neural network to classify images as damage and no-damage	Ecuador Earthquake	Mathew Hurricane	68.7	79.1	68.3	72.6
		Mathew Hurricane	Ruby Typhoon	68.1	66.9	77.4	71.3
		Nepal Earthquake	Mathew Hurricane	70.6	86.0	63.4	72.4
[26]	Unsupervised DA using Maximum Mean Discrepancy metric to classify image tweets	Hurricane Harvey	Six disasters (Average of all Six)	79.33	78.23	78.09	77.75
-	Baseline Model (SSDA): Single-Source DA for image classification	Hurricane Harvey	Hurricane Maria	76.43	76.61	76.47	76.41
		Hurricane Harvey	Iraq-Iran Earthquake	72.95	71.12	73.48	71.22
		Hurricane Harvey	California Wildfires	78.57	77.10	77.00	77.17
-	Baseline Model (CSDA): Combined-Source DA for the classification of image tweets	Hurricane Harvey + Hurricane Irma	Hurricane Maria	77.31	77.36	77.29	77.29
		Hurricane Harvey + Hurricane Irma	Iraq-Iran Earthquake	73.77	74.16	77.29	73.03
		Hurricane Harvey + Mexico Earthquake	Iraq-Iran Earthquake	82.79	80.61	82.71	81.18
		Hurricane Harvey + Mexico Earthquake	California Wildfires	78.57	77.48	78.36	77.77
-	Proposed Model (MSDA-DM): Multi-Source DA for the binary classification of image tweets: ‘informative’ and ‘non-informative’	Hurricane Harvey + Hurricane Irma	Hurricane Maria	79.08	79.66	79.04	78.95
		Hurricane Harvey + Hurricane Irma	Iraq-Iran Earthquake	79.51	77.43	80.27	78.09
		Hurricane Harvey + Mexico Earthquake	Iraq-Iran Earthquake	83.61	81.47	83.32	82.05
		Hurricane Harvey + Mexico Earthquake	California Wildfires	72.05	70.99	71.85	71.55