Add topics to a topic model behind your text prediction by uploading a file with
the topics and their associated training data for machine learning models, or keywords
for keyword-based models. Templates are available for download to help you create a
topic file more easily.
Tip:
Machine learning is a more effective method of
training topic models than keywords. Keywords are Boolean matches that
cannot identify the most accurate topic. Do not use keywords in
production scenarios.
Keywords influence the behavior of a machine learning model, but they
are not exact rules. The Should, Must, and And
words act as positive features for matching a text to a topic, while the
Not words act as negative features. The training and testing data have
the greatest impact on your machine learning model, while keywords have a smaller
impact.
Open the text prediction:
In the navigation pane of
App Studio, click Channels.
In the Current channel interfaces section, click
the icon that represents a channel for which you want to configure the
text prediction.
On the channel configuration page, click the
Behavior tab, and then click Open
text prediction.
In the Prediction workspace, click OutcomesTopics.
In the Language field, select the language for which you
want to add a topic.
Result: The change of language refreshes the list of available topics. The list
now displays the topics for the selected language.
Click Import, and then select one of the following
options:
To upload topics with their respective training data, click
Machine learning.
To upload topics (taxonomy) with their respective keywords, click
Keywords.
In the Topic model field, select the topic model to
which you want to add the topics.
If you do not have a topic file ready, create a topic file by using the
template:
In the Topic template field, click
Download to download the topic
template.
Fill in the topic template with the information for the topics that you
want to upload.
For example: A topic file for a machine learning model can contain the following topics
and training data:
content
result
type
I want to book a flight
ticket
action > book ticket
What is the price for a ticket from London to
Dubai
action > book ticket
I want to take a trip to
Tokyo
action > book ticket
test
I want to cancel my ticket
action > cancel ticket
can you help me cancel my
reservation
action > cancel ticket
cancel PNR number 27382
action > cancel ticket
test
You can mark some pieces of training data as test data. The system
uses the training data to train the model and the test data to test the
model and generate its f-score. If you do not specify any test data, the system
randomly assigns a portion of the training data as test data, typically,
using the 70:30 ratio of training data to test data.
Add the topic file:
Click Choose File, and then browse for the topic
file.
Select the file, and then click Open.
Click Upload.
Result: The topics are added to the
Topics list with pending training data. You can use
this training data directly for model building.
In the prediction workspace, click Save to save your
changes.
For example:
To learn how to import topics to a text prediction, see the following
video: