Normalize

Use Normalize seed with --minMax. Normalize one or more columns using a min max scaler

Examples

Example 1 - Normalize Entire Dataframe with Min-Max Scaling

In many modeling tasks, it’s important to normalize all numeric values so they are on the same scale. This example normalizes every numeric column in the dataframe using min-max scaling, which scales values to fall between 0 and 1.

#> Normalize --all --minMax
AFLEFT 
minMaxScaler = MinMaxScaler()
for column in appleStockDf.columns:
    if is_numeric_dtype(appleStockDf[column]):
        appleStockDf[column] = minMaxScaler.fit_transform(appleStockDf[ [column] ]) AFRIGHT

Example 2 - Normalize Selected Columns with Min-Max Scaling

Rather than normalizing every column, you can choose specific columns to scale. In this case, we normalize only the High and Low columns using min-max scaling, which helps maintain focus on features of interest.

#> Normalize --columns High Low --minMax
AFLEFT 
scaleColumn = appleStockDf[ ['High', 'Low'] ]
scaleColumn = minMaxScaler.fit_transform(scaleColumn)
appleStockDf[ ['High', 'Low'] ] = scaleColumn AFRIGHT

Example 3 - Normalize Columns Conditionally

Sometimes normalization should only apply to a subset of the data. This example applies min-max scaling to the Open column, but only for rows where the value in the High column is greater than 100. All other rows remain unchanged.

#> Normalize --columns Open --minMax --where High > 100
AFLEFT 
scaleColumn = appleStockDf[ ['Open'] ][appleStockDf['High'] > 100]
scaleColumn = minMaxScaler.fit_transform(scaleColumn)
appleStockDf.loc[appleStockDf['High'] > 100,'Open'] = scaleColumn AFRIGHT