Description of the final prediction system » Historie » Verze 1
Roman Kalivoda, 2021-06-14 09:03
| 1 | 1 | Roman Kalivoda | h1. Description of the final prediction system |
|---|---|---|---|
| 2 | |||
| 3 | The final prediction system consists of an arbitrary number of classifiers. The number and structure of classifiers can be configured in @<path_to_executable>/../Predictor.config@ file by altering the total number of predictors used and assigning the building areas to groups. |
||
| 4 | Another way to configure the predictions is to change the implementation of the @Predictor/Ipredictor.class@ interface. There are currently two pre-created implementations, NaiveBayesClassifier and SdcaMaximumEntropyClassifier, which have better results. |