Cleaning your data before building your Time Series model is crucial.
Learn how to do it, step by step 🧵👇
8
68
309
Replies
1️⃣ Handle missing values
2️⃣ Remove trend
👇
1
1
4
3️⃣ Remove seasonality
4️⃣ Check for stationarity and make it stationary if necessary
5️⃣ Normalize the data
👇
2
1
4
6️⃣ Remove outliers
👇
1
1
3
7️⃣ Smooth the data
👇
1
1
4
You may change this order depending on your data. Or even remove some steps or repeat some of them several times on different data cleaning stages.
This has the sole purpose of showing you one possibility.
I hope it was useful! 😉
0
1
4
@daansan_ml
Why do we have to remove seasonality? Can't we use that as a feature in our ML model?
0
0
0
@daansan_ml
Last question. If you have a stationary mean and stationary variance based off the data, how would you fix this? I guess my question is why do we have to check for stationarity, and what do we do to address the issue?
0
0
1
@opinionsplainer
@daansan_ml
Saved! Here's the compiled thread:
🪄 AI-generated summary:
"Cleaning data before building a Time Series model is essential. This thread provides a step-by-step guide on how to do it, including handling missing values, removing trend,...
1
0
1