I have a numpy array
z = array(['Iris-setosa', 'Iris-setosa', 'Iris-setosa', 'Iris-setosa','Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor','Iris-virginica', 'Iris-virginica', 'Iris-virginica'])
I want to replace
Iris-setosa -0
Iris-versicolor - 1
Iris-virginica - 2
to apply logistic regression.
Final output should be like
z = [ 0, 0 ,.. 1,1,.. 2,2,..]
Is there a simple way to do this operation instead of iterating through the array and use replace command?
Use factorize:
a = pd.factorize(z)[0].tolist()
print (a)
[0, 0, 0, 0, 1, 1, 1, 2, 2, 2]
Or numpy.unique:
a = np.unique(z, return_inverse=True)[1].tolist()
print (a)
[0, 0, 0, 0, 1, 1, 1, 2, 2, 2]
you can use a dictionary:
my_dict = {'Iris-setosa': 0, 'Iris-versicolor': 1, 'Iris-virginica': 2}
then use list comprehension:
z = [my_dict[zi] for zi in z]
Are you trying to count the number of occurrence as you are trying to do logistic regression?
If yes, you can use the following as well.
import collections
z = ['Iris-setosa', 'Iris-setosa', 'Iris-setosa', 'Iris-setosa','Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor','Iris-virginica', 'Iris-virginica', 'Iris-virginica']
print (collections.Counter(z))
It will print as below:
Counter({'Iris-setosa': 4, 'Iris-versicolor': 3, 'Iris-virginica': 3})
If you want to print in another way, you can do the following:
import collections
z = ['Iris-setosa', 'Iris-setosa', 'Iris-setosa', 'Iris-setosa','Iris-versicolor', 'Iris-versicolor', 'Iris-versicolor','Iris-virginica', 'Iris-virginica', 'Iris-virginica']
for item in collections.Counter(z):
print(str(item)+ ' ' + str(collections.Counter(z)[item]))
The output will be
Iris-setosa 4
Iris-versicolor 3
Iris-virginica 3
[list(set(z)).index(val) for val in z]
simply put, cast a set out of z to get only unique values, then cast a list out of that set for indexing, then finally use a list comprehension to get the final list. If you have a very large list of strings, I would suggest setting list(set(z)) to a variable outside of the list comprehension
[2, 2, 2, 2, 0, 0, 0, 1, 1, 1, 2] but isn't Iris-setosa be set to 0
[list(np.unique(z)).index(val) for val in z]
pd.Series(z, dtype="category"), see pandas.pydata.org/pandas-docs/stable/categorical.htmlmachine-learningquestion; please do not spam the tag (removed)