How to evaluate a dataset in KnowledgeFlow with a previously saved model
I have a model generated with a numerical dataset in Explorer and I want to use this model to evaluate new datasets in KnowledgeFlow. The final idea is automate a process, using this KnowledgeFlow in a server, where new numerical datasets enter, the model is applied in these datasets, and new results datasets are saved in.
So, what is the process to evaluate a dataset in KnowledgeFlow with a previously saved model?
Re: How to evaluate a dataset in KnowledgeFlow with a previously saved model
Grab an appropriate data source from the “DataSources” menu. Connect this source to a “ClassAssigner” component from the “Evaluation” menu. Use a “dataSet” connection for this. Then, connect the “ClassAssigner” component to the “AddClassification” component (from “Filters -> supervised -> attribute”), again using a “dataSet” connection. Finally, connect the “AddClassification” component to an appropriate data sink component from the “DataSinks” menu.
Configure the sink and the source to read and write respectively to appropriately chosen files. Configure “ClassAssigner” so that the class attribute is set correctly. Configure “AddClassification” to read the “serializedClassifierFile” and set “outputClassification” to “True”. You may also want to set “removeOldClass” to “True”.