Trend Analysis on Time Series Data

Hi,

I am trying to include a trend on my time series data. Let’s say I have monthly data and I need to get a trend whether the value stucks, rises or falls on a statistical basis.

E. g.:
Month Value
1 10
2 11
3 8
4 15
5 18
6 20
7 22
8 23

I would now need to calculate for months 6, 7 and 8 for example the statistical trend.

I am not sure I am understanding. Are you trying to see whether the value for this month is higher than the value for the previous month or whether the value inside the month is growing?
From the data up here it seems like the first option.

Rosaria

Hi again,

in principal I need this:
https://hub.knime.com/knime/space/Examples/50_Applications/26_Model_Process_Management/Process/Products/workflows/Predict_upOrDown/04_Learn_Trend

Saying this I have problems to run this example since the table file is missing (anyone knows where I can get this)?

Furthermore I discussed this example a bit with our data scientist (he isn’t experienced in Knime) and we setup a possible flow for my data. Now I have to transfer it to my entire data set (I extracted from my dataset one group but I have to loop over all groups). Maybe you have some directions for that as well? I mean with group that I want to have a trend for each group in my data. A group consists out of 6 time series samples and I have several groups(categories)…

Hi there,

have you had any progress with this issue?

What exactly you need from attached workflow? Perform linear regression on your data or something else? The attached workflow is part of model factory ecosystem so I would guess that the table file missing is coming from model factory connection. Have you checked the whitepaper for more information? https://files.knime.com/sites/default/files/inline-images/Model_Process_Management_20170404_1.pdf

Br,
Ivan

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.