Fig. 13: “Sampling-feature-hyperparameters-weights” simultaneous optimization process. | npj Digital Medicine

Fig. 13: “Sampling-feature-hyperparameters-weights” simultaneous optimization process.

From: Integrating Internet multisource big data to predict the occurrence and development of COVID-19 cryptic transmission

Fig. 13

a ICHIO consists of a dataset sampling strategy, feature probabilities, feature scales, hyperparameters and weights. b Comparison of feature probabilities (Fm) and feature scales (Cc) for each base learner with feature subsets selected depending on the size of the relationship. c The result of feature selection. d Combination of “sampling-feature-hyperparameter-weight” and simultaneous optimization to obtain the optimal individual value after several iterations of the objective function.

Back to article page