Fig 1.
Overview of the synthetic data utilized in the study.
A: The blue frame represents the data generation step, the green frame signifies data pre-processing, and the yellow frame indicates data partitioning. B: Examples of simulation from the SIR model with white noise.
Fig 2.
32 classifiers trained from every unique combination of four synthetic training datasets, two feature extraction libraries, and four predictive models.
The initial letter in each classifier’s name corresponds to the training set (W, WhiteN; E, EnvN; D, DemN; M, MixedN). The middle number represents the feature extraction (22, 22 statistical features; 5, 5 early warning signal indicators). The last letter represents the predictive model (G, GBM; L, LRM; K, KNN; S, SVM).
Fig 3.
The AUC values of 32 synthetic-data-trained classifiers (horizontal axis, see Fig 2) on withheld testing sets.
Classifiers are reordered by AUC scores. Error bars correspond to the 95% confidence intervals. DeLong tests are conducted to compare the AUC values of classifiers, with detailed results available in Tables A-D in S2 Text (where the predictive model is fixed) and Tables E-L in S2 Text (where the training data is fixed).
Fig 4.
Illustration in Rolling window experiment settings and AUC score at different reiterations.
A: Rolling window settings on and
. We maintain a fixed length of input time series L and roll the window L from the right to the left boundary. B-I: AUC scores depicting the classification performance of the classifier trained from changing gaps D between the subsequence and transition points. Two feature extraction methods (22SF, 22 statistical features; 5EWSI, 5 early warning signal indicators) are put in two rows and four predictive models (GBM, gradient boosting machine; LRM, logistic regression model; KNN, k-nearest neighbor; SVM, support vector machine) are presented in four columns.
Fig 5.
Illustration in Expanding window experiment settings and AUC score at different reiterations.
A: Expanding window settings on and
. We fix the length of D as 30, then expand L until reaching the left boundary of sliced data. The initial length of L is 5. B-I: AUC scores depicting the classification performance of the classifier trained from changing gaps D between the subsequence and transition points. Two feature extraction methods (22SF, 22 statistical features; 5EWSI, 5 early warning signal indicators) are put in two rows and four predictive models (GBM, gradient boosting machine; LRM, logistic regression model; KNN, k-nearest neighbor; SVM, support vector machine) are presented in four columns.
Fig 6.
Classification accuracy of 32 synthetic-data-trained classifiers (horizontal axis, see Fig 2) to correctly classify subsequences of four empirical datasets.
Error bars correspond to the 95% confidence intervals.