Hi Heather,
Thank you for your question. At the moment, the outlier detection algorithm is done automatically when we are building our forecast models for ABM method (which will normalize outliers when generating forecast), while for weighted historical method, you would need to manually weigh past historical data to generate your forecast (i.e., you can weigh or exclude outlier days to discount them). The ABM method looks at past historical data up to 5 years and automatically seeks to detect outliers and repeatable events when projecting forward. It would also detect long-term trend and seasonality and gives you the best forecast that validates well against historical data. I would definitely suggest you to try it out if you haven't already.
Also FYI, the feature to mark historical data as outliers or special events (repeated or one-off) is currently in our roadmap and will be implemented as part of our continuous forecasting feature, which will inform our ABM algorithm to take into account from your input. Hope this helps.
------------------------------
Bayu Wicaksono
Genesys - Employees
------------------------------
Original Message:
Sent: 03-03-2023 10:44
From: Heather Higginbotham
Subject: Smoothing a forecast
Is there a way to mark in genesys intervals that we want to exclude from the forecast due to an outage or some anomaly during the interval that throws off the numbers?
#Reporting/Analytics
------------------------------
Heather Higginbotham
Cox Automotive Corporate Services, LLC
------------------------------