Creative Commons source reading · Prediction and forecasting
A novel lightweight deep learning model for early prediction of cardiovascular disease
Nasir Gul; Aurangzeb Khan; Mazliham Mohd Su'ud; Muhammad Mansoor Alam; Danyal Skandri
Originally published in Front Artif Intell · 2026-05-20
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
Cardiovascular Disease (CVD) is still the main cause of death globally. Hence, to have timely clinical intervention, there is a need for the prediction of early risk models which are accurate and dependable. Here is a lightweight and strong artificial intelligence-based system for the early prediction of CVD using clinical data. We have created an end-to-end method comprising data preprocessing, stratified train and test splitting, and class imbalance treatment by SMOTE exclusively on the training set to ensure that there is no data leakage. We evaluate a series of machine learning (ML) and deep learning (DL) models, including Logistic Regression (LR), Decision Tree (DT), K-Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machine (SVM), and various Artificial Neural Network (ANN) architectures, systematically. In order to…
Results and experiments
The goal of this research is to develop a lightweight deep learning approach for precisely and timely recognition of CVD’s initial indications in order to manage the anomaly before it causes major consequences. The study used both machine learning and deep learning approaches, which are described in detail below, along with validation findings.
Discussion
Several machine learning techniques, including LR, SVM, KNN, RF, and DT, were utilized in the proposed research to detect early CVD symptoms. RF outperformed other ML algorithms with an accuracy of 98.50%. Before training, the collection of data was balanced with SMOTE to minimize bias towards the larger group, as the initial collection had 480 normal instances and 520 CVD instances. The major goal of this work was to apply and test DL models, which are recognized for producing more broad outcomes than ML models, whereas ML algorithms were utilized as a baseline method. The proposed study evaluated a number of DL architectures, including ANNs, RNNs, LSTMs, and 1D-CNNs. In comparison to ML, DL-based approaches have been described as data-hungry; hence, the training set used in this study was of a medium size and was effectively…
Editorial reading guide
- What data created and tested the model, and was evaluation independent?
- What is the cost of a false positive, false negative or out-of-distribution decision?
