More detailed drug information is available in these columns. These columns include information on drug categories, which is coded. We must omit these columns because they represent future information. However, this information could be immensely useful in other machine learning problems:
drug_id_cols = [
'DRUGID1','DRUGID2','DRUGID3','DRUGID4','DRUGID5',
'DRUGID6','DRUGID7','DRUGID8','DRUGID9','DRUGID10',
'DRUGID11','DRUGID12'
]
drug_lev1_cols = [
'RX1V1C1','RX1V1C2','RX1V1C3','RX1V1C4',
'RX2V1C1','RX2V1C2','RX2V1C3','RX2V1C4',
'RX3V1C1','RX3V1C2','RX3V1C3','RX3V1C4',
'RX4V1C1','RX4V1C2','RX4V1C3','RX4V1C4',
'RX5V1C1','RX5V1C2','RX5V1C3','RX5V1C4',
'RX6V1C1','RX6V1C2','RX6V1C3','RX6V1C4',
'RX7V1C1','RX7V1C2','RX7V1C3','RX7V1C4',
'RX8V1C1','RX8V1C2','RX8V1C3','RX8V1C4',
'RX9V1C1','RX9V1C2','RX9V1C3','RX9V1C4',
'RX10V1C1','RX10V1C2','RX10V1C3','RX10V1C4',
'RX11V1C1','RX11V1C2','RX11V1C3','RX11V1C4',
'RX12V1C1','RX12V1C2','RX12V1C3','RX12V1C4'
]
drug_lev2_cols = [
'RX1V2C1','RX1V2C2','RX1V2C3','RX1V2C4',
'RX2V2C1','RX2V2C2','RX2V2C3','RX2V2C4',
'RX3V2C1','RX3V2C2','RX3V2C3','RX3V2C4',
'RX4V2C1','RX4V2C2','RX4V2C3','RX4V2C4',
'RX5V2C1','RX5V2C2','RX5V2C3','RX5V2C4',
'RX6V2C1','RX6V2C2','RX6V2C3','RX6V2C4',
'RX7V2C1','RX7V2C2','RX7V2C3','RX7V2C4',
'RX8V2C1','RX8V2C2','RX8V2C3','RX8V2C4',
'RX9V2C1','RX9V2C2','RX9V2C3','RX9V2C4',
'RX10V2C1','RX10V2C2','RX10V2C3','RX10V2C4',
'RX11V2C1','RX11V2C2','RX11V2C3','RX11V2C4',
'RX12V2C1','RX12V2C2','RX12V2C3','RX12V2C4'
]
drug_lev3_cols = [
'RX1V3C1','RX1V3C2','RX1V3C3','RX1V3C4',
'RX2V3C1','RX2V3C2','RX2V3C3','RX2V3C4',
'RX3V3C1','RX3V3C2','RX3V3C3','RX3V3C4',
'RX4V3C1','RX4V3C2','RX4V3C3','RX4V3C4',
'RX5V3C1','RX5V3C2','RX5V3C3','RX5V3C4',
'RX6V3C1','RX6V3C2','RX6V3C3','RX6V3C4',
'RX7V3C1','RX7V3C2','RX7V3C3','RX7V3C4',
'RX8V3C1','RX8V3C2','RX8V3C3','RX8V3C4',
'RX9V3C1','RX9V3C2','RX9V3C3','RX9V3C4',
'RX10V3C1','RX10V3C2','RX10V3C3','RX10V3C4',
'RX11V3C1','RX11V3C2','RX11V3C3','RX11V3C4',
'RX12V3C1','RX12V3C2','RX12V3C3','RX12V3C4'
]
addl_drug_cols = [
'PRESCR1','CONTSUB1','COMSTAT1','RX1CAT1','RX1CAT2',
'RX1CAT3','RX1CAT4','PRESCR2','CONTSUB2','COMSTAT2',
'RX2CAT1','RX2CAT2','RX2CAT3','RX2CAT4','PRESCR3','CONTSUB3',
'COMSTAT3','RX3CAT1','RX3CAT2','RX3CAT3','RX3CAT4','PRESCR4',
'CONTSUB4','COMSTAT4','RX4CAT1','RX4CAT2','RX4CAT3',
'RX4CAT4','PRESCR5','CONTSUB5','COMSTAT5','RX5CAT1',
'RX5CAT2','RX5CAT3','RX5CAT4','PRESCR6','CONTSUB6',
'COMSTAT6','RX6CAT1','RX6CAT2','RX6CAT3','RX6CAT4','PRESCR7',
'CONTSUB7','COMSTAT7','RX7CAT1','RX7CAT2','RX7CAT3',
'RX7CAT4','PRESCR8','CONTSUB8','COMSTAT8','RX8CAT1',
'RX8CAT2','RX8CAT3','RX8CAT4','PRESCR9','CONTSUB9',
'COMSTAT9','RX9CAT1','RX9CAT2','RX9CAT3','RX9CAT4',
'PRESCR10','CONTSUB10','COMSTAT10','RX10CAT1','RX10CAT2',
'RX10CAT3','RX10CAT4','PRESCR11','CONTSUB11','COMSTAT11',
'RX11CAT1','RX11CAT2','RX11CAT3','RX11CAT4','PRESCR12',
'CONTSUB12','COMSTAT12','RX12CAT1','RX12CAT2','RX12CAT3',
'RX12CAT4'
]
X_train.drop(drug_id_cols, axis=1, inplace=True)
X_train.drop(drug_lev1_cols, axis=1, inplace=True)
X_train.drop(drug_lev2_cols, axis=1, inplace=True)
X_train.drop(drug_lev3_cols, axis=1, inplace=True)
X_train.drop(addl_drug_cols, axis=1, inplace=True)
X_test.drop(drug_id_cols, axis=1, inplace=True)
X_test.drop(drug_lev1_cols, axis=1, inplace=True)
X_test.drop(drug_lev2_cols, axis=1, inplace=True)
X_test.drop(drug_lev3_cols, axis=1, inplace=True)
X_test.drop(addl_drug_cols, axis=1, inplace=True)