Figure 4
The multi-panel layout shows seven scatter plots arranged across two rows, each panel containing a square grid with both axes ranging from 0.0 to 1.0 with an interval of 0.1 and a diagonal reference line from the lower-left corner to the upper-right corner. Blue circular points are plotted in each panel. On the left side of every panel, a control section includes labels “Y Axis: C S C” and “X Axis: Model underscore 1”, “Model underscore 2”, “Model underscore 3”, “Model underscore 4”, “Model underscore 5”, “Model underscore 6”, and “Model underscore 7”, along with “Case I D Column”, a “Plot” button, and a “Zoom” checkbox. Below these, two values are shown as “Consistency X less than or equal to Y” and “Consistency X greater than or equal to Y”. In the first panel labeled “A D asterisk tilde C C asterisk tilde C L asterisk tilde I O T asterisk tilde I S asterisk tilde S G O” with “X Axis: Model underscore 1”, most points are concentrated between x equals 0.0 and 0.2, forming a dense vertical band, with vertical values ranging from about 0.05 up to about 0.95. A few points appear between x equals 0.4 and 0.6 at mid and lower vertical positions. The values shown are “Consistency X less than or equal to Y: 0.843949” and “Consistency X greater than or equal to Y: 0.214055”. In the second panel labeled “A D asterisk C C asterisk tilde C L asterisk tilde I O T asterisk tilde I S asterisk S G O” with “X Axis: Model underscore 2”, points cluster below x equals 0.2 with vertical values from about 0.05 to 0.95, and several points extend toward x equals 0.4 to 0.6. The values shown are “Consistency X less than or equal to Y: 0.855742” and “Consistency X greater than or equal to Y: 0.246769”. In the third panel labeled “tilde A D asterisk C C asterisk tilde C L asterisk tilde I O T asterisk tilde I S asterisk tilde P R asterisk tilde S G O” with “X Axis: Model underscore 3”, points are distributed between x equals 0.0 and 0.4 with vertical spread from about 0.05 to 0.95, and a few points appear near x equals 0.5. The values shown are “Consistency X less than or equal to Y: 0.878924” and “Consistency X greater than or equal to Y: 0.1979”. In the fourth panel labeled “A D asterisk C C asterisk tilde C L asterisk tilde I O L asterisk I S asterisk tilde asterisk tilde S G O” with “X Axis: Model underscore 4”, a dense cluster is visible near x less than 0.2 with vertical values extending from about 0.05 to near 1.0, and additional points appear between x equals 0.3 and 0.5. The values shown are “Consistency X less than or equal to Y: 0.869318” and “Consistency X greater than or equal to Y: 0.216277”. In the fifth panel labeled “tilde A D asterisk C C asterisk tilde C L asterisk I O T asterisk I S asterisk tilde P R asterisk tilde S G O” with “X Axis: Model underscore 5”, points cluster between x equals 0.0 and 0.3 with vertical values from about 0.05 to 0.95, and a few points extend toward x equals 0.5 at lower and mid vertical levels. The values shown are “Consistency X less than or equal to Y: 0.81698” and “Consistency X greater than or equal to Y: 0.237076”. In the sixth panel labeled “tilde A D asterisk C C asterisk tilde C L asterisk tilde I O T asterisk I S asterisk P R asterisk tilde S G O” with “X Axis: Model underscore 6”, most points lie below x equals 0.2 with vertical spread from about 0.05 to 0.95, and scattered points appear near x equals 0.3 to 0.5. The values shown are “Consistency X less than or equal to Y: 0.879381” and “Consistency X greater than or equal to Y: 0.217892”. In the seventh panel labeled “A D asterisk tilde C C asterisk tilde C L asterisk I O T asterisk tilde I S asterisk tilde P R asterisk S G O” with “X Axis: Model underscore 7”, points cluster near x less than 0.3 with vertical values extending from about 0.05 to 0.95, and a few points appear between x equals 0.4 and 0.6. The values shown are “Consistency X less than or equal to Y: 0.795377” and “Consistency X greater than or equal to Y: 0.236269”. Note: All numerical data values are approximated.Fuzzy-plot of models for low CSC (from Table 9) using data from the holdout sample