Predict air quality with machine learning code issue

[joshuazchambers]

When I run the code provided with the experiment my MAE is 13.---- as I was reading trying to find out what that really meant, that is basically the amount of units the prediction is off by, but in the explanation it say that if it is about .13 its off by 13 units, I am just asking if I am doing the code wrong or is the explanation worded wrong, should my MAE be closer to .13? I am having a similar issue with the MSE coming out to 387.----. Please get back to me soon. Thank you.



[Administrator: Project url: https://www.sciencebuddies.org/science- ... ir-quality ]

[amyCC]

Hi - Thank you for asking about this. It sounds like you are in good shape with your numbers. I think the example of .13 in the project may be a bit misleading. (It is being updated.)

According to the project creator, the unit being predicted in this project is AQI. Given that, an error of 13 (what you got for the MAE) is not out of scale. The scientist notes: "There is not much of a difference between an AQI of 100 and an AQI 113."

The MSE (Mean Squared Error) seems very large because the error is squared. If you take the square root of your MSE, it is relatively close to your MAE.

Let us know if you have other questions.

Amy
Science Buddies