Research Article
Heart Risk Failure Prediction Using a Novel Feature Selection Method for Feature Refinement and Neural Network for Classification
Table 5
Classification accuracies of the proposed method and other methods in the literature that used heart disease dataset.
| Study (year) | Method | Accuracy (%) |
| ToolDiag, RA [35] | IB1-4 | 50.00 | WEKA, RA [35] | InductH | 58.50 | ToolDiag, RA [35] | RBF | 60.00 | WEKA, RA [35] | FOIL | 64.00 | ToolDiag, RA [35] | MLP + BP | 65.60 | WEKA, RA [35] | T2 | 68.10 | WEKA, RA [35] | 1R | 71.40 | WEKA, RA [35] | IB1c | 74.00 | WEKA, RA [35] | | 76.70 | Robert Detrano [35] | Logistic regression | 77.00 | Paul et al. [19] | FDSS | 80 | Khemphila and Boonijing (2011) [36] | MLP-backpropagation | 80.99 | Newton Cheung (2001) [37] | BNNF | 80.96 | Newton Cheung (2001) [37] | C4.5 | 81.11 | Newton Cheung (2001) [37] | Naïve Bayes | 81.48 | Newton Cheung (2001) [37] | BNND | 81.11 | WEKA, RA [35] | Naïve Bayes | 83.60 | Šter and Dobnikar [38] | Fisher discriminant analysis | 84.2 | Šter and Dobnikar [38] | Linear discriminant analysis | 84.5 | Šter and Dobnikar [38] | Naïve Bayes | 82.5–83.4 | Polat et al. (2005) [25] | AIRS | 84.50 | Dwivedi (2018) [22] | LR | 85 | Ozsen et al. (2005) [39] | Kernel functions with AIS | 85.93 | Kahramanli and Allahverdi (2008) [40] | Hybrid neural network system | 86.8 | Polat et al. (2006) [10] | Fuzzy-AIRS-knn based system | 87.00 | Amin et al.(2019) [23] | LR, SVM, K-NN | 87.4 | Verma et al. (2016) [20] | CFS-PSO | 88.4 | Das et al. (2009) [1] | Neural network ensembles | 89.01 | Jankowski and Kadirkamanathan (1997) [41] | IncNet | 90.00 | Ali and Bukhari (2019) [12] | Mutual information + DNN | 90 | Kumar (2011) [42] | ANFIS | 91.18 | Shah et al. (2017) [21] | PPCA | 91.30 | Samuel et al. (2017) [4] | ANN-fuzzy-AHP | 91.10 | Kumar (2012) [43] | Fuzzy resolution mechanism | 91.83 | Ali et al. (2019) [34] | Stacked SVMs | 92.22 | Proposed method (2020) | Feature selection based on FWAFE + ANN | 91.11 | Proposed method (2020) | Feature selection based on FWAFE + DNN | 93.33 |
|
|