Select the prediction dataset on this screen. Let’s review the prediction data used in this tutorial.

Use the prediction model you have created to predict the complaint type for review statements for which you have not yet labeled the complaint type. In this tutorial, you will use sample data that has been prepared for prediction.

Specify the prediction data 2_CustomerFeedback(For Prediction).csv here.

Click [Select from Uploaded Data] and select the sample data from the data list on the [Samples] tab.

The prediction dataset preview is displayed. Predict “Failure Type (target)” displayed as “?” from information such as “Contents of the post”, and “Review”.

Click [Run Prediction]. Wait a while until the preview screen of the prediction result is displayed.

Please specify “Customer ID” in [Add the following variables to the first column]. Click [Save Prediction Results], enter “File name” and click [Save].

When the prediction is complete, the following screen is displayed and the prediction results are saved in the specified file.

Predicted results are output in the following format (this format may vary depending on the option settings). The probability of predicting which complaint type each customer feedback is classified as is calculated.