Skip to main content

Table 5 LDA Geometric shapes

From: Deriving image features for autonomous classification from time-series recurrence plots

STANDARD Circle Ellipse Rectangle Square Triangle Prediction success
Circle 50 0 0 0 0 1.00
Ellipse 0 45 0 5 0 0.90
Rectangle 0 0 50 0 0 1.00
Square 1 6 1 37 5 0.74
Triangle 0 0 7 4 39 0.78
Model       0.884
RQA       
Circle 47 0 3 0 0 0.94
Ellipse 0 41 0 6 3 0.82
Rectangle 1 0 48 1 0 0.96
Square 1 10 6 31 2 0.62
Triangle 1 5 0 2 42 0.84
Model       0.836
STANDARD & RQA       
Circle 50 0 0 0 0 1.00
Ellipse 0 46 0 4 0 0.94
Rectangle 0 0 50 0 0 1.00
Square 1 1 2 43 3 0.88
Triangle 0 0 0 0 50 1.00
Model       0.956
STANDARD & RQA & ESQ       
Circle 50 0 0 0 0 1.00
Ellipse 0 46 0 4 0 0.92
Rectangle 0 4 49 1 0 0.90
Square 1 3 1 43 2 0.86
Triangle 0 0 0 0 50 1.00
Model       0.952
  1. Confusion matrices of the different parameter combinations used for the LDA analyses of geometric shapes
  2. Values in bold represent true positive classifications of the LDA model and overall discriminative success