ROLLME: readme and choice node

Author: matnar

examples/sentiment-analysis/README.md

@@ -15,7 +15,8 @@ The Sentiment Analysis (SA) workflow combines function tasks and choice tasks.
 SA consists of the following tasks:
 * RetrieveState (`sa_retrieve`): Retrieves the dataset;
 * ExtractState (`sa_extract`): Preprocess the dataset;
-* ChoiceState: Choose whether to train and test either a low or a high accuracy model; 
+* ChoiceState: Choose whether to train and test either a low or a high accuracy model;
+  If the input parameter `max_features` is below 10000, the low-accuracy model will be used in the remainder of the workflow;
 * LATrainState (`sa_train`): The training task of the low-accuracy sentiment analysis model;
 * LAEvaluateFinalState (`sa_evaluate`): The final task of the low-accuracy sentiment analysis model;
 * HATrainState (`sa_train`): The training task of the high-accuracy sentiment analysis model;