Ensemble Models Clause Samples

Ensemble Models. For every possible trial in Section 4.1, 4.2, 4.3, further ensemble models are used. Ensemble models are using different sub-models as the machine learning method to learn different information, and combine the final results from every sub-model. Ensemble Model 1: Select one best model on work experience, education and work experience and education, respectively(total 3 models), then com- bine their prediction result, and vote for the final result. Ensemble Model 2: From all models performing work experience, educa- tion and work experience and education, select top 3 models for each(total 9 models), then combine three models prediction result for each part, and vote for the final result. Ensemble Model 3: Combine the 9 models in ensemble 3, aggregate the prediction result and vote for the final prediction result.