The problem here is doing House Price Prediction when it came, try to do GarageArea log transformation, but due to some houses no garage, so GarageArea = 0, then became -inf by log transformation, so I want all - inf data is converted back to all 0, as shown in the title.
In addition to recycling can also how to do it? Look at the following code
train_test['GarageArea'] = np.log(train_test['GarageArea']) train_test['GarageArea'] = train_test['GarageArea'].replace(float("-inf"), 0)
Originally thought replace () function is a specific function str, try to discover the results can also be used in the array, where special -inf is written, can not be directly written -inf or '-inf', and to write float ( "- inf ")