How do we standardize the data in the predictive tools?

[Gualigee]

Hi 

For instance, we have temperatures, soil moisture, and precipitation predictor variables. They have different units of measure.

How do we appropriately standardize predictor variables for models other than linear regression, logistic regression, and neural networks in Alteryx?

In addition, if I have used the PCA tool to reduce the no. of variables for temperatures, soil moisture, and precipitation (climate variables have 12 months' data), do I need to consider standardization again in the predictive tool configuration? I mean I have standardized the 12 variables for temperatures, soil moisture, and precipitation respectively in the PCA tool. 

Thank you very much. 

[IraWatt · Accepted solution]

Hey @Gualigee,

I believe you should scale the data before PCA machine learning - Do you standardize the data before PCA. There's a great article here on how to standardise data in Alteryx Normalization, Standardization, and Regularization... - Alteryx Community It covers PCA, standardization and links to a lot of other great work. 

[Gualigee]

Hi @IraWatt, thank you for the article recommended. 

[Gualigee]

@IraWatt the principal component tool has a check box to scale the data. so we can standardize the data and reduce dimensionality at the same time?