Preprocessing the Discrete Variables

In the lesson Data Preparation. Preprocessing Discrete Variables: Creating dummies ,  For the variable 'addr_state' it was explained that grouping into categories for the final model has to be done according to the WoE values keeping into the consideration of the No.of observations(Borrowers) But , in the homework problem , when we try to preprocess the discrete variable 'purpose' shouldn't we group them just as above , then into the following categories: # small_business, educational, moving ,house
# Other
# renewable_energy, medical, wedding, vacation
# debt_conslidation
#home_improvement
#major_purchase, car
#credit_card But the solution is given as : # We combine 'educational', 'small_business', 'wedding', 'renewable_energy', 'moving', 'house' in one category: 'educ__sm_b__wedd__ren_en__mov__house'.
# We combine 'other', 'medical', 'vacation' in one category: 'oth__med__vacation'.
# We combine 'major_purchase', 'car', 'home_improvement' in one category: 'major_purch__car__home_impr'.
# We leave 'debt_consolidtion' in a separate category.
# We leave 'credit_card' in a separate category. Can you just explain me on what basis the above grouping is being done?