[Paper Review] A Benchmark Study by using various Machine Learning Models for Predicting Covid-19 trends

D. Kamelesun, R. Saranya+1 more|arXiv (Cornell University)|Jan 26, 2023

COVID-19 diagnosis using AIMedicine3 citations

TL;DR

This benchmark study evaluates multiple machine learning models—linear and polynomial regression, KNN, SVM, decision trees, random forests, ensemble methods (voting, bagging, stacking), and deep learning (ANN)—on real-world COVID-19 datasets to predict future trends in confirmed cases. Random Forest and ensemble models outperformed others, while SVR and Lasso regression showed poor performance due to difficulty in separating complex, non-linear patterns in the data.

ABSTRACT

Machine learning and deep learning play vital roles in predicting diseases in the medical field. Machine learning algorithms are widely classified as supervised, unsupervised, and reinforcement learning. This paper contains a detailed description of our experimental research work in that we used a supervised machine-learning algorithm to build our model for outbreaks of the novel Coronavirus that has spread over the whole world and caused many deaths, which is one of the most disastrous Pandemics in the history of the world. The people suffered physically and economically to survive in this lockdown. This work aims to understand better how machine learning, ensemble, and deep learning models work and are implemented in the real dataset. In our work, we are going to analyze the current trend or pattern of the coronavirus and then predict the further future of the covid-19 confirmed cases or new cases by training the past Covid-19 dataset by using the machine learning algorithm such as Linear Regression, Polynomial Regression, K-nearest neighbor, Decision Tree, Support Vector Machine and Random forest algorithm are used to train the model. The decision tree and the Random Forest algorithm perform better than SVR in this work. The performance of SVR and lasso regression are low in all prediction areas Because the SVR is challenging to separate the data using the hyperplane for this type of problem. So SVR mostly gives a lower performance in this problem. Ensemble (Voting, Bagging, and Stacking) and deep learning models(ANN) also predict well. After the prediction, we evaluated the model using MAE, MSE, RMSE, and MAPE. This work aims to find the trend/pattern of the covid-19.

Motivation & Objective

• To evaluate the effectiveness of diverse machine learning models in predicting future trends of COVID-19 confirmed cases.
• To compare traditional supervised learning models (e.g., linear regression, SVM, decision trees) with ensemble and deep learning approaches.
• To identify which models perform best on real-world, non-linear, time-series pandemic data.
• To assess model performance using standard regression metrics (MAE, MSE, RMSE, MAPE) across different prediction scenarios.

Proposed method

• The study uses historical COVID-19 case data as input for training supervised machine learning models.
• Models evaluated include linear regression, polynomial regression, K-nearest neighbors (KNN), support vector regression (SVR), decision trees, random forests, and ensemble techniques (voting, bagging, stacking).
• A feedforward artificial neural network (ANN) is also implemented as a deep learning baseline.
• Model training and prediction are performed on time-series data of confirmed cases, with hyperparameters tuned for optimal performance.
• Performance is quantitatively evaluated using MAE, MSE, RMSE, and MAPE to compare predictive accuracy across models.
• The benchmark includes both individual model evaluation and ensemble stacking to improve generalization and robustness.

Experimental results

Research questions

• RQ1 Which machine learning model yields the most accurate predictions for future COVID-19 confirmed case counts across diverse datasets?
• RQ2 How do traditional models like SVR and Lasso regression compare to ensemble and deep learning models in capturing non-linear trends in pandemic data?
• RQ3 To what extent do ensemble techniques (e.g., bagging, stacking) improve prediction performance over individual models?
• RQ4 Why do SVR and Lasso regression underperform in this specific time-series forecasting task?

Key findings

• Random Forest and decision tree models achieved the highest predictive accuracy among all models tested.
• Ensemble methods such as voting, bagging, and stacking demonstrated strong performance, outperforming individual models.
• Support Vector Regression (SVR) showed consistently low performance due to difficulty in separating complex, non-linear patterns in the data.
• Lasso regression also exhibited poor performance, likely due to its sensitivity to high-dimensional or noisy time-series data.
• Artificial Neural Networks (ANN) performed well, indicating potential for deep learning in pandemic forecasting.
• The evaluation metrics (MAE, MSE, RMSE, MAPE) confirmed that Random Forest and ensemble models had the lowest error rates across all metrics.