If I am having a variable with two values (for example Sex can take male or female), I use code like,
train_df["Sex"] = train_df["Sex"].apply(lambda sex: 0 if sex == 'male' else 1)
to convert string to integer. What is the way to do it if the variable takes more than 2 values, like Salary categorised as low/medium/high? How to assign value similarly as above?
Use map by dictionary:
d = {
'male': 0,
'female': 1,
'other': 2
}
train_df["Sex"] = train_df["Sex"].map(d)
But for Salary is better cut if need new values by ranges:
train_df = pd.DataFrame({'Salary': [100,200,300,500]})
bins = [0, 200, 400, np.inf]
labels=['low','medium','high']
train_df['label'] = pd.cut(train_df['Salary'], bins=bins, labels=labels)
print (train_df)
Salary label
0 100 low
1 200 low
2 300 medium
3 500 high
You can make a transformation dict for example:
values = {
"low" : 0,
"med" : 1,
"high": 2
}
train_df["Sex"] = train_df["Sex"].apply(lambda level: values.get(level, 0))
