[BUG] NaiveEnsembleModel predictions have wrong component names

[DavidKleindienst]

Describe the bug
Predictions should have the same component names as the TimeSeries used to fit the model.
For NaiveEnsembleModel.predict that's not the case - The component of the returned TimeSeries is always called "0"

To Reproduce

import pandas as pd
import numpy as np
from darts import TimeSeries
from darts.models import NaiveEnsembleModel, NaiveMean

dates = pd.date_range("2020-01-01", "2020-12-31", freq = "D")
nr_dates = len(dates)
ts=TimeSeries.from_times_and_values(dates, 3*np.ones(nr_dates), columns=["TS"])
model = NaiveEnsembleModel([NaiveMean()])
model.fit(ts)
prediction = model.predict(10)
assert ts.components == prediction.components

raises an AssertionError

Expected behavior
No assertion error should be raised

System (please complete the following information):

• Python version: 3.10
• darts version 0.23.1