Icon

01_​Emil_​the_​TeacherBot

Display MostProbable PredictedClasses Emil the TeacherBot - Deployment via WebPortal InterfaceThis workflow deploys a teacher bot. That is:- it displays a textbox for the user to ask a question - it parses the text in the question to extract keywords- it queries a machine learning model to predict the question class- it asks the user to select one among the three most probably classes- it then retrieves the top resources links to tutorials) available on the KNIME web site for that category.- it ends with a Goodbye pageClasses "Bug", "Legal", "Announcement", or "Something else", forward the question to a KNIME expert.The page with the links to tutorials also includes a Send Email button as emergency exit.The Goodbye page updates the model training set with the question and the selected class. If class "Something Else" was selected, the question gets manually relabelled for the training set. Thank You pages Suggest Resources Concatenate KNIMEResources andNode Guide Process the Question,Predict the Classes,and Get the two mostProbable Ones User Input - GatherQuestion from theUser Update Datasets Summary of the Question& Question Action required!Select aclass in View. document class predictionQueries a ML modeldisplay list of linksto suggested tutorialsFinal ClassFollow ClassPathFoundLinks?Find appropriatelinks to tutorialsUser Input onthe WebPortal Choose Answer- Deployment Get_Probabilities Suggest_Resources IF Switch Happy Ending End-KNIMESupportTeam End IF End IF IF Switch End-Special Empty Table Switch End IF No Links Available KNIME Resources Update Datasets Display MostProbable PredictedClasses Emil the TeacherBot - Deployment via WebPortal InterfaceThis workflow deploys a teacher bot. That is:- it displays a textbox for the user to ask a question - it parses the text in the question to extract keywords- it queries a machine learning model to predict the question class- it asks the user to select one among the three most probably classes- it then retrieves the top resources links to tutorials) available on the KNIME web site for that category.- it ends with a Goodbye pageClasses "Bug", "Legal", "Announcement", or "Something else", forward the question to a KNIME expert.The page with the links to tutorials also includes a Send Email button as emergency exit.The Goodbye page updates the model training set with the question and the selected class. If class "Something Else" was selected, the question gets manually relabelled for the training set. Thank You pages Suggest Resources Concatenate KNIMEResources andNode Guide Process the Question,Predict the Classes,and Get the two mostProbable Ones User Input - GatherQuestion from theUser Update Datasets Summary of the Question& Question Action required!Select aclass in View. document class predictionQueries a ML modeldisplay list of linksto suggested tutorialsFinal ClassFollow ClassPathFoundLinks?Find appropriatelinks to tutorialsUser Input onthe WebPortal Choose Answer- Deployment Get_Probabilities Suggest_Resources IF Switch Happy Ending End-KNIMESupportTeam End IF End IF IF Switch End-Special Empty Table Switch End IF No Links Available KNIME Resources Update Datasets

Nodes

Extensions

Links