Sorry- this coder has been taken down due to server abuse. Please contact michael.tanana@utah.edu for more info.

This coder is designed to take one utterance at a time. (An utterance is usually defined as a single statement or idea, and is often about the length of a single sentence.)

Enter the utterance into the textbox and click on whether it should be coded as a client statement or a counselor statement.

If the model does not correctly classify the statement in your opinion, you may click on the correct code at the bottom of the page under "Correct Model"

Example Client Utterance: "something happened in the motel where I said I just can't do this anymore"

Example Counselor Utterance: "you said that you graduated and that you did accomplish some of your goals it sounds like"

  

Classname

Correct Model

Class Probabilities

Explain Models

DSF: Discrete Sentence Features These are features from the utterance that are coded as 0's and 1's. These include n-grams (uni-bi-tri) and features from the Stanford Dependency Parse

RNN: Recursive Neural Network This is a variation on a recursive neural network that uses the structure of the dependency parse tree to create a continuous (real valued) representation of the sentence.

About