This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
Export a Trained Topic Classification Model to Categorize New Items
Taken from giphy.com
Once we create and train our model to identify topics within the data we have, the next step is to use that trained model to assign topics to new data we receive and thus avoid running the entire preparation and training process all over again, which would consume more time and resources.
The image above is an end-to-end topic identification process. How can we make it more efficient?
The first step is to place an Output Data tool after the M anchor of the Topic Modeling tool.
Select an Alteryx Database file (.yxdb) as the output format, as this will store the model object, and we can use it to assign topics based on the data we used to train it. Having the model trained, we can use it in a new workflow.
As demonstrated in the above image, place the new dataset that we want to classify (Text Input connected to a Text Pre-processing tool) and the trained model that we exported (Input Data tool with TrainedTopicModeling.yxdb). To score the new values, we will use the Predict Values tool found in the Machine Learning tab of the tool palette.
Connect the trained model to input anchor M (model), and the data to anchor D (data). It is important to add text preprocessing first to prepare the new data correctly.
And that's it! When executing the workflow, your model will assign a topic to each data point. This way we will optimize the execution time and we can use the model to categorize new data easily.