Data Science for Marketing Analytics, Second Edition by Mirza Rahim Baig | Gururajan Govindan | Vishwesh Ravi Shrimali

Author:Mirza Rahim Baig | Gururajan Govindan | Vishwesh Ravi Shrimali
Language: eng
Format: epub, mobi
Publisher: Packt Publishing Pvt. Ltd.
Published: 2021-09-06T00:00:00+00:00

Mean Absolute Error

There are multiple ways to use residuals to evaluate a model. One way is to simply take the absolute value of all residuals and calculate the average. This is called the Mean Absolute Error (MAE), and can intuitively be thought of as the average difference you should expect between your model's predictions and the true value:

Figure 6.3: Equation for calculating the MAE of a model

In the preceding equation, yj is the true value of the outcome variable for the data point j, and Å·j is the prediction of the model for that data point. By subtracting these terms and taking the absolute value, we get the residual. n is the number of data points, and thus by summing over all data points and dividing by n, we get the mean of the absolute error.

Therefore, a value of zero would mean that your model predicts everything perfectly, and larger values mean a less accurate model. If we have multiple models, we can look at the MAE and prefer the model with the lower value.

We can implement MAE using the scikit-learn library, as shown here:

We will import the mean_absolute_error metric from the sklearn.metrics module:from sklearn.metrics import mean_absolute_error

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.