cancel
Showing results for 
Search instead for 
Did you mean: 

Believability of Bins in Time Series

Believability of Bins in Time Series

Please take a look at the following snapshot for a pretty basic model that I ran earlier today:

zsfeinsteinyahoocom_0-1657591373857.png

This looks pretty darn good given it is quite a Univariate test with a prediction going out by one day

Can anyone comment on the model comparison that I did here?

Or how about the following Accuracy over time?

zsfeinsteinyahoocom_1-1657591575783.png

Am still learning the intricacies of the DataRobot models

I next need to see how I may export the results for my one-day prediction

Thank you!

0 Kudos
2 Replies
dalilaB
DataRobot Alumni

For accuracy and consistent prediction, the AVG Blender is sligthly better then xGboost, but not significantly, when looking a the lift chart.  xGboost lift chart is not monotonically increasing.  However, the slightly better accuracy has a cost.  AVG Blender is almost twice as slow as xGboost.   

zsfeinstein@yahoo.com   Your model comparison is of the two models against each other (yellow and blue lines) and with the Dual Lift you could see how they perform against actual values as well. From the MASE scores both of the models are about 14%-15% better than the Baseline or naiive model. 

The DataRobot University instructor-led Time Series Modeling course ( https://university.datarobot.com/datarobot-time-series-modeling ) provides good examples of how to develop and make predictions with the platform.

0 Kudos