A Deep Learning–Based Predictive Model for Myocardial Infarction Risk Forecasting Using Patient Comorbidities and Lifestyle Data

PID	Gender	Age (years)	Stress level	Blood sugar	Resp. rate (breaths/min)	Heart rate (bpm)	Blood pressure (SBP/DBP)	BMI status	Comorbidities	Lifestyle factors
2	M	77	High	Elevated	22	63	126/84	Normal	Heart disease, diabetes	Non-smoker, healthy diet
7	M	43	High	Normal	20	81	118/84	Obese	Heart disease, insomnia	Smoker, junk food
9	M	50	Medium	Normal	18	83	111/84	Normal	None	Smoker, junk food
17	M	23	Medium	Normal	19	83	113/86	Normal	Sleep apnea	Healthy food, active
21	M	55	Low	Normal	20	85	126/85	Normal	None	Non-smoker, active
1	F	37	High	Elevated	23	60	124/86	Over weight	Stress disorder	Smoker, junk food
4	F	41	High	Normal	22	99	118/72	Obese	Sleep apnea	Sedentary, junk food
5	F	25	Medium	Normal	21	69	138/76	Obese	Sleep apnea	Smoker, junk food
8	F	72	Low	Normal	18	96	135/77	Normal	None	Healthy food, active
15	F	89	Low	Normal	17	98	127/87	Normal	None	Non-smoker, balanced

[i] BMI, body mass index; DBP, diastolic blood pressure; MI, myocardial infarction; PID, patient identification number; SBP, systolic blood pressure.

**(A)** Evaluation of MI risk with the execution of SVM. **(B)** MI risk with the execution of SMOTE. ELU, exponential linear unit; MI, myocardial infarction; NB, Naïve Bayes; RF, random forest; SHAP, Shapley additive explanations; SMOTE, synthetic minority oversampling technique; SVM, support vector machine.

Table 2:

Decision boundary comparison.

Method	Samples properties	Result
SVM-based: Imbalanced data	Normal samples: N N N N N N N N N N N N N N N N N Anomaly samples: A A	Boundary optimized for the majority class Anomalies misclassified as normal (high false negatives—dangerous for medical diagnosis)
SMOTE-based: Balanced data	Normal samples: N N N N N N N N N N N N Risk samples: R R R R R R R R R R R R	Boundary fairly separates both classes Better classification of the minority class (Better sensitivity—critical for early detection)

[i] SMOTE, synthetic minority oversampling technique; SVM, support vector machine.

Distribution plot of patients' vitals. BMI, body mass index; VLDL, very low-density lipoprotein.

Correlation heat map for vitals. BMI, body mass index; VLDL, very low-density lipoprotein.

Confusion matrix for models TabNet, RF, NB. NB, Naïve Bayes; RF, random forest.

**(A)** Feature impact on model output. **(B)** Feature impact on model output. SHAP, Shapley additive explanations.

Table 4:

Model performance of RF.

	Precision	Recall	F1 score	Support
0	0.94	1.00	0.97	5,553
1	0.82	0.21	0.33	447
Accuracy			0.94	6,000
Macro avg	0.88	0.60	0.65	6,000
Weighted avg	0.93	0.94	0.92	6,000

[i] RF, random forest.

Table 5:

Model performance of NB.

	Precision	Recall	F1 score	Support
0	0.95	0.98	0.96	5,553
1	0.53	0.33	0.41	447
Accuracy			0.93	6,000
Macro avg	0.74	0.65	0.68	6,000
Weighted avg	0.92	0.93	0.92	6,000

[i] NB, Naïve Bayes.

Table 6:

Model performance of TabNet.

	Precision	Recall	F1 score	Support
0	0.94	0.98	0.96	5,553
1	0.57	0.27	0.36	447
Accuracy			0.93	6,000
Macro avg	0.76	0.63	0.66	6,000
Weighted avg	0.92	0.93	0.92	6,000

[i] Remember: RF (0.21)> TabNet (0.27)> NB (0.33)

[ii] Precision: TabNet (0.57)> RF (0.82)> NB (0.53)

ROC curve comparison of OC curve comparison of TabNet, and RF. NB, Naïve Bayes; RF, random forest.

**(A)** ROC curve comparison of the enhancing approach. **(B)** Usage of ELU in the feature extraction layer prior to the RF. ELU, exponential linear unit; MI, myocardial infarction; NB, Naïve Bayes; RF, random forest.

Table 7:

Model performance of ELU on RF.

	Precision	Recall	F1 score	Support
Class 0	0.98	0.99	0.99	5,553
Class 1	0.90	0.85	0.88	447
Accuracy	-	-	0.97	6,000
Macro avg	0.94	0.92	0.93	6,000
Weighted avg	0.97	0.97	0.97	6,000

[i] ELU, exponential linear unit; RF, random forest.

Table 8A:

Data summary.

Summary	Values
Total_samples	308,854
Total_columns	40
Total_attributes	399
Training_sample	306,186
Test_sample	92,657
Target_column	Risk
Target_classes	2
Class_training	{0:153093,1:153093}
Class_test	{0:65612,1:27045}

Table 8B:

Classification report.

	Precision	Recall	F1 score	Support
Class 0	1.00	0.99	1.00	65,612
Class 1	0.98	1.00	0.99	27,045
Accuracy	-	-	0.99	92,657
Macro avg	0.99	1.00	0.99	92,657
Weighted avg	0.99	0.99	0.99	92,657

Investigation of feature importance related to the feature vector. BMI, body mass index; RF, random forest.

Confusion matrix of RF following SMOTE. RF, random forest; SMOTE, synthetic minority oversampling technique.

Comparison of improved accuracy and AUC. ELU, exponential linear unit; RF, random forest; SMOTE, synthetic minority oversampling technique; SVM, support vector machine.

Metrics comparison of RF. RF, random forest; SMOTE, synthetic minority oversampling technique; SVM, support vector machine.

Table 9:

Classification report of ELU on RF.

	Precision	Recall	F1 score	Support
Class 0	1.00	0.99	1.00	65,612
Class 1	0.98	0.99	0.99	27,045
Accuracy	-	-	0.99	92,657
Macro avg	0.99	1.00	0.99	92,657
Weighted avg	0.99	0.99	0.99	92,657

**(A)** SHAP for two dominant features. **(B)** Mean SHAP for features. BMI, body mass index; SHAP, Shapley additive explanations; VLDL, very low-density lipoprotein.

Table 10:

Samples and their properties.

Samples	Properties
Sample 1	Sample index: 82,496; true label: 1 (Class 1); predicted probabilities: [0.26460299 0.73539701]; predicted class: 1 (Class 1)
Sample 2	Sample index: 51,516; true label: 0 (Class 0); predicted probabilities: [0.60369369 0.39630631]; predicted class: 0 (Class 0)
Sample 3	Sample index: 55,756; true label: 0 (Class 0); predicted probabilities: [0.624137 0.375863]; predicted class: 0 (Class 0)

LIME analysis for sample 2 (Class 0). BMI, body mass index.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.2478/ijssis-2026-0044 | Journal eISSN: 1178-5608

Journal RSS Feed

Language: English

Submitted on: Sep 11, 2025

Published on: Jul 18, 2026

Published by: International Journal on Smart Sensing and Intelligent Systems

In partnership with: Paradigm Publishing Services

Publication frequency: Volume open

Keywords:

myocardial infarction,

ELU,

SMOTE,

SHAP,

LIME

Related subjects:

Engineering,

Introductions and overviews,

Engineering, other

© 2026 T. Sarkar, B. Kundu, E. Chakraborty, D. Ganguly, Sarin Raj, published by International Journal on Smart Sensing and Intelligent Systems
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Volume 19 (2026): Issue 1 (January 2026)

A Deep Learning–Based Predictive Model for Myocardial Infarction Risk Forecasting Using Patient Comorbidities and Lifestyle Data

Figures & Tables

Table 1:

Figure 1:

Table 2:

Figure 2:

Figure 3:

Figure 4:

Figure 5:

Table 4:

Table 5:

Table 6:

Figure 6:

Figure 7:

Table 7:

Table 8A:

Table 8B:

Figure 8:

Figure 9:

Figure 10:

Figure 11:

Table 9:

Figure 12:

Table 10:

Figure 13:

Figure 14:

Figure 15:

Paradigm

My account