| Sumario: | Diagnosing heart disease is considered a difficult task as it provides a digital estimation of the seriousness of the disease. As a result, the quickest treatment can be done. So, heart diagnosis has attracted a lot more attention in the medical industry throughout the globe, and along with excellence in efficacy, the optimization algorithm plays a crucial role in the detection of heart disease. Here, to predict the cardiac illness an improvised CatBoost algorithm and Multi-layer Perceptron classifier are used. Also, proper hyperparameter tweaking is needed for the successful implementation of the classifier. In order to optimize the hybrid model's hyperparameters, the Mayfly optimization algorithm is deployed for effective hyperparameter optimization. In order to increase prediction accuracy, the Harris-Hawks optimization technique is used to choose the essential features from the dataset. Z-Alizadeh Sani and Cleveland heart disease datasets are utilized to detect heart disease. Also, it is compared with the existing models. To validate the efficiency of a model, six various measures are used: precision, accuracy, recall, the F- 1 measure, specificity, and loss. Here, when compared to the previous studies, the proposed model yields better performance, i.e., 98.7% accuracy with Cleveland and 99.2% with Alizadeh Sani Datasets.
|
|---|