Figure 1 Distance Matrix of Height and Weight Data.

Figure 4 Percent of Each Variable Entry With Missing Data in the Myocardial Infarct Dataset.

Figure 7 Comparison of Classifier Performance. Left: Estimated of Accuracy Scores for KNN, Decision Tree. Right: Neural Network Classifiers. These Measures are Taken From 5-Fold Cross-Validation for Each Model.

Figure 8 Scree Plot Displaying the Percentage of Variance Captured by the First 25 Principal Components of the Myocardial Infraction Dataset. Sometimes Referred to as an Elbow Plot, a Vertical Line Has Been Drawn at PC6, the Point at Which the Additional Variance Captured Visibly Decreases or “Elbows”.

Figure 11 UMAP Embeddings of the Myocardial Infarction Dataset, Generated Using Different min_dist Values. Points are Colored by LET_IS Value.

Figure 2 The Distances Among Samples from the Myocardial Infarction Dataset are Shown According to Different Distance Metrics. While Qualitatively, These Matrices Look Similar, Note That the Scale of the Distances Varies Greatly Between Metrics.

Figure 3 Scatter Plot of Systolic Blood Pressures Recorded by the Intensive Care Unit and Emergency Cardiology Team. The First Principal Component (Red) and the Second Principal Component (Blue) are Also Shown.

Figure 6 Visualization of KNN Versus RKNN for k = 1. Given a Query Point q , KNN Finds the Point Closest to q ( ), While RKNN Identifies All Points in the Dataset for Which q is a KNN. The Latter can be Achieved by Drawing a Circle Around Each Point with a Radius to its -Nearest Neighbor. q is