The Prediction Algorithm and Characteristics Analysis of Kuroshio Sea Surface Temperature Anomalies
Table 2
Diagnostic rule set of whether Spring Kuroshio SST is high discovered by CHAID algorithm.
Decision rules
Whether the SST is high
Decision attributes
Learning accuracy
Rule A: If North American polar vortex intensity index ≤2992.602
YES
North American polar vortex intensity index
2/2 = 100.00%
Rule B: If North American polar vortex intensity index >2992.602 and W-shaped index of North Atlantic-European circulation >24.333
YES
North American polar vortex intensity index, W-shaped index of North Atlantic-European circulation
2/3 = 66.67%
Rule C: If North American polar vortex intensity index >2992.602 and W-shaped index of North Atlantic-European circulation ≤24.333 and north Africa-North Atlantic-North America subtropical high area index ≤2.62
NO
North American polar vortex intensity index, W-shaped index of North Atlantic-European circulation, North Africa-North Atlantic-North America subtropical high area index
20/20 = 100.00%
Rule D: If North American polar vortex intensity index >2992.602 and W-shaped index of North Atlantic-European circulation ≤24.333 and (North Africa-North Atlantic-North America subtropical high area index >2.62 and North Africa-North Atlantic-North America subtropical high area index ≤3.803)
YES
North American polar vortex intensity index, W-shaped index of North Atlantic-European circulation, North Africa-North Atlantic-North America subtropical high area index
2/2 = 100.00%
Rule E: If North American polar vortex intensity index >2992.602 and W-shaped index of North Atlantic-European circulation ≤24.333 and North Africa-North Atlantic-North America subtropical high area index >3.803
NO
North American polar vortex intensity index, W-shaped index of North Atlantic-European circulation, North Africa-North Atlantic-North America subtropical high area index