Exponentially weighted running mean (moving average) using Pandas

Asked May 02 '19 at 15:28

Active May 02 '19 at 15:28

Viewed 117 times

I have been trying to get the exponentially weighted running mean for a series of mean daily outdoor temperatures (temp_series), defined as follows:

I would like to use a value of alpha=0.8 and input a custom period of days. I have tried using:

trm = temp_series.ewm(alpha=0.8).mean()

and

trm = df_wea_dd.ewm(alpha=0.8, min_periods=30, adjust=False).mean()

but the values I obtain are not reliable (I have checked using excel for a period of 30 days). What am I doing wrong?

asked May 02 '19 at 15:28

Andreuccio

1,053
2
18
32

What do you mean by "reliable"? – doctorlove May 02 '19 at 15:30
`ewm().mean()` computes running mean over the entire history. what do you mean by *a custom period of days*. – Quang Hoang May 02 '19 at 16:23
you should provide a [mcve] to remove as much ambiguity as possible – piRSquared May 02 '19 at 16:26
@doctorlove it means I am getting results that are significantly different from those I obtain using excel for validation – Andreuccio May 08 '19 at 17:19
@piRSquared you are right, my question could be better framed. I ll provide an example – Andreuccio May 08 '19 at 17:20
Did you manage to find an answer to this? I am guessing you are trying to calculate tc on CIBSEs model. I have been stuck in the same place. @Andreuccio – rohit Dec 15 '22 at 15:21

Exponentially weighted running mean (moving average) using Pandas

0 Answers0