Do you need to make the data stationary for deep ar?

To prepare time series data for forecasting, some common preprocessing steps include:

• Checking for and handling missing data. This could involve imputing missing values.
• Making the time series stationary. Many forecasting models assume the time series is stationary. This means the statistical properties like mean and variance do not change over time. Differencing or detrending can help make the series sttionary.
• Feature engineering. Deriving additional features from the raw time series can help improve forecast accuracy. These features could include things like seasonality patterns, price effects, promotional periods etc.
• Data transformation and scaling. Since different algorithms work best with data scaled in a particular range, it's often necessary to transform and scale variables. For example, standardizing data to have mean of 0 and standard deviation of 1.
• Train-test split. The data should be divided into training and test sets to develop and evaluate the model performance. A common split is 70% for training and 30% for testing.