Skip to main content

Task 3: Set up experiment

Now that we have a better understanding of the dataset, let's build the experiment from scratch.

  1. In the DATASETS page, click on the consumer_complaint_resolution_train.csv dataset, and select the PREDICT option.

  2. As soon as you select the Predict option, you are asked if you want to take a tour of the Driverless AI environment. Skip it for now by clicking Not Now.

  3. Feed the following information into Driverless AI: name-experiment

    note

    a. Display Name - Name the current experiment: tutorial-4c.

    b. Target Column - Select Consumer disputed? as the target column. The aim of the experiment is to predict whether a customer will dispute the resolution of their complaint. The column has only two values, yes and no.

    c. Dropped Columns - For this experiment, the primary focus is on text columns. Therefore, drop all columns that are not relevant for text analysis, such as IDs, date or other non-informative columns. Please note that if you decide to keep the non-text columns, the NLP algorithms will still work on the non-text columns. Dropped columns

    d. Test Dataset - The Test dataset is a dataset used to provide an unbiased evaluation of a final model fit on the training dataset. It is not used during training of the model. Select the consumer_complaint_resolution_test.csv dataset for the test dataset option.

In Task 4: Configure training settings, let's continue editing the experiment settings.

Feedback