Hyperparameter optimization for cardiovascular disease data-driven prognostic system

Abstract

Prediction and diagnosis of cardiovascular diseases (CVDs) based, among other things, on medical examinations and patient symptoms are the biggest challenges in medicine. About 17.9 million people die from CVDs annually, accounting for 31% of all deaths worldwide. With a timely prognosis and thorough consideration of the patient's medical history and lifestyle, it is possible to predict CVDs and take preventive measures to eliminate or control this life-threatening disease. In this study, we used various patient datasets from a major hospital in the United States as prognostic factors for CVD. The data was obtained by monitoring a total of 918 patients whose criteria for adults were 28-77 years old. In this study, we present a data mining modeling approach to analyze the performance, classification accuracy and number of clusters on Cardiovascular Disease Prognostic datasets in unsupervised machine learning (ML) using the Orange data mining software. Various techniques are then used to classify the model parameters, such as k-nearest neighbors, support vector machine, random forest, artificial neural network (ANN), naïve bayes, logistic regression, stochastic gradient descent (SGD), and AdaBoost. To determine the number of clusters, various unsupervised ML clustering methods were used, such as k-means, hierarchical, and density-based spatial clustering of applications with noise clustering. The results showed that the best model performance analysis and classification accuracy were SGD and ANN, both of which had a high score of 0.900 on Cardiovascular Disease Prognostic datasets. Based on the results of most clustering methods, such as k-means and hierarchical clustering, Cardiovascular Disease Prognostic datasets can be divided into two clusters. The prognostic accuracy of CVD depends on the accuracy of the proposed model in determining the diagnostic model. The more accurate the model, the better it can predict which patients are at risk for CVD.

Keywords: Cardiovascular disease; Data mining; Data-driven analytics; Hyperparameter optimization; Orange data mining software; Prognostic system; Unsupervised machine learning.

Fig. 1

CVD mortality rates are expected to increase significantly by 2020 [3]

Fig. 2

Trends in CVD mortality for men and women in United States from 1980 to 2019 [3]

Fig. 3

Research flowchart

Fig. 4

Classification techniques used to classify Cardiovascular Disease Prognostic datasets

Fig. 5

Test and score results

Fig. 6

Classification accuracy matrix

Fig. 7

Calibration plot based on classification accuracy CVD prognostic. a target = 0; b target = 1

Fig. 8

F1 score matrix

Fig. 9

Precision matrix

Fig. 10

Recall matrix

Fig. 11

Calibration plot based on F1 score CVD prognostic. a target = 0; b target = 1

Fig. 12

Calibration plot based on precision and recall CVD prognostic. a target = 0; b target = 1

Fig. 13

Performance curve analysis CVD prognostic. a target = 0; b target = 1

Fig. 14

Clustering techniques are used to analyze Cardiovascular Disease Prognostic datasets

Fig. 15

K-means clustering

Fig. 16

K-means clustering silhouette scores

Fig. 17

Scatter plot of k-means cluster analysis. a K-means clustering scatter plot correlations between cholesterol level and maximum HR attributes; b K-means clustering scatter plot correlations between cholesterol level and resting blood pressure attributes; c K-means clustering scatter plot correlations between maximum HR and resting blood pressure attributes

Fig. 18

Distances

Fig. 19

Hierarchical clustering. a Hierarchical clustering based on the cholesterol level attribute; b Hierarchical clustering based on the maximum HR attribute; c Hierarchical clustering based on the resting blood pressure attribute

Fig. 20

Hierarchical clustering silhouette scores

Fig. 21

Hierarchical cluster scatter diagram that illustrates the correlation between cholesterol level, maximum HR, and resting blood pressure. a Hierarchical clustering scatter plot correlations between cholesterol level and maximum HR attributes; b Hierarchical clustering scatter plot correlations between cholesterol level and resting blood pressure attributes; c Hierarchical clustering scatter plot correlations between maximum HR and resting blood pressure attributes

Fig. 22

DBSCAN clustering

Fig. 23

DBSCAN clustering silhouette scores

Fig. 24

The scatter plot of DBSCAN cluster that demonstrates the correlation among cholesterol level, maximum HR, and resting blood pressure. a DBSCAN clustering scatter plot correlations between cholesterol level and maximum HR attributes; b DBSCAN clustering scatter plot correlations between cholesterol level and resting blood pressure attributes; c DBSCAN clustering scatter plot correlations between maximum HR and resting blood pressure attributes