Edit Column Details
When building your model, it's important to ensure that the variable metadata is correct and update the column details where required.
Key things to check and update, where required, are:
- Edit the data Description, if required.
- Ensure that the Data Type and Statistical Type are correct.
- Smart Predict provides an initial guess of the data type and statistical type, but this can be incorrect. To optimize the automated data encoding, the correct data and statistical types are essential.
- If the guessed data or statistical type is not correct, you must update it before training your model.
- In the Interpret As Missing Value column enter a missing value description if there are missing values.
- Enter the string that is used to represent the missing values, for example, 999, #NA or #Empty. This value is now assigned as missing in the data encoding process and is not considered to be an actual data value.
- For example, if a value of 999 is entered into the Interpret As Missing description for a continuous type variable, then the data encoding will assign any 999 values as missing.
- Indicate which variables are Keys.