Interaction terms

[Hoeze]

Hi, I would like to test the significance of a variable with an interaction term as covariate: trait ~ sex:age + variable_to_test Can I do this with your library @jzluo?

[jzluo]

Hi @Hoeze, this would have to done in the data preparation just like sklearn. Say you have a dataframe df with columns trait, sex, age, variable_to_test. Two simple examples:

• Use patsy

  y, X = dmatrices('trait ~ sex:age + variable_to_test', df)

• Create interaction term in pandas

  df['sex:age'] = df['sex'] * df['age']
  X, y = df[['sex:age', 'variable_to_test']], df['trait']

Let me know if it works for you!