Research Article
Creating Ensemble Classifiers with Information Entropy Diversity Measure
Algorithm 1
Incremental_SEM algorithm.
| Input: | Training dataset with labels by ensemble classifier Lt−f; | The interval threshold of classification diversity: [a, b]; | Iterate number (each iterate creates a new base classifier: k) | Ensemble classifier at period of [t−f, t]: Lt−f; | Output: incremental ensemble classifier Lt at time t. | (1) | Begin | (2) | Loop | (3) | Compute diversity value λ0 of ensemble classifier Lt−f; | (4) | If | (5) | For i = 1 to k | (6) | Sampling training data from labeled dataset at period of [t−f, t] by Lt−f; | (7) | Generate a new base classifier Li; | (8) | Add Li to Lt−f; | (9) | Compute the diversity value λ1; | (10) | If | (11) | Lt = Lt−f ; | (12) | Return Lt | (13) | End for | (14) | else if | (15) | Compute the accuracy of each base classifier at Lt−f; | (16) | Sort base classifiers in decreasing order of accuracy as baselist; | (17) | Delete some member base classifiers with the lowest accuracy at Lt−f; | (18) | Update the Lt−f; | (19) | Lt = Lt−f; | (20) | return Lt; | (21) | else | (22) | Lt = Lt−f; | (23) | return Lt | (24) | End if | (25) | Break; | (26) | End loop |
|