Optimizing Deep Learning Model for Software Cost Estimation Using Hybrid Meta-Heuristic Algorithmic Approach

Table 3

Comparison between existing approaches.

Refer ence	DL/ML/ANN	Meta-heuristic algorithm	Dataset	Evaluation parameter	Contributions
[42]	NN	Fiery algorithm, BAT algorithm	COCOMO81, NASA, MAXWELL, China	MRE, MMRE, pred, MDMRE	Hybrid model for effort estimation
[45]	ANN	Firefly	COCOMO81, NASA, MAXWELL, China	MMRE, MdMRE, PRED	Hybrid model for cost estimation
[46]	DNN	Cuckoo, hybrid PSO	COCOMO	RE, MRE, MMRE, MARE, PRED, execution time	Hybrid model for cost estimation
[11]	DNN	GA	KC1, KC2, CM1, PC1, JM1	Accuracy, precision, F-score, recall, sensitivity	Defect prediction
[47]	DL	Evolutionary algorithm	KEEL dataset repository	Accuracy, G-mean, precision, F-score, computational time	Hybrid of DBN and ADE for imbalanced classification
[48]	NN	GA, PSO	N/A	Survey	Possibility to apply on DL
[49]	NN	Cuckoo	COCOMO	MMRE, standard deviation	Improve cocomo
[50]	ANN	Cuckoo	COCOMO81, NASA	MMRE, PRED, computational time	Hybrid model
[51]	GWO, HSA	SBA	NASA	MRE, MMRE	Hybrid model