Table 3.

Previous studies related to heart disease.

Reference	Approach	Dataset	Challenges
56	Multi-Kernel Learning and Neuro-Fuzzy Inference	KEGG Metabolic Dataset	inappropriate performance evaluation metrics
60	BiGRU, BiLSTM, CNN ensemble	Cardiac Disease Dataset	The study uses the dataset is not a standard dataset
61	RF, XGB, and GBC	Multiple heart disease datasets such as Cleveland, Hungarian, Z-Alizadeh Sani Dataset	Stacking three models introduces higher complexity and cost to the model
62	Fisher technique and Generalized Discriminant Analysis	two datasets NSR-CAD and SR-CAD	The training process on large datasets for heart rate variability needs improvement due to the existing lack of training
63	NB, k-NN, RF, and DT	Multiple heart disease datasets such as Cleveland, Hungary, VA Long Beach, Switzerland Datasets	It is important to conduct further experimentation with different combinations of models and feature choices to explore potential enhancements
64	LSTM and RNN	IoT cloud data from UCI machine learning repository’s Cleveland and Hungarian datasets	The challenges include managing vast amounts of patient data, integrating Internet of Things (IoT) devices, utilizing deep learning for accurate disease diagnosis, ensuring data quality, addressing privacy and security concerns, achieving early disease recognition, and extracting meaningful features from the data
65	Bi-LSTM	data collected from IoT sensors, electronic clinical data (ECD)	Key challenges include managing and processing vast healthcare data, ensuring data quality, maintaining privacy and security, early disease detection, and personalized healthcare