Скачать или смотреть How to Unnormalize Data Using MinMaxScaler in Python

Learn how to easily `unnormalize` your data using MinMaxScaler in Python to obtain original values from normalized predictions.
---
This video is based on the question https://stackoverflow.com/q/66952880/ asked by the user 'sera' ( https://stackoverflow.com/u/15522353/ ) and on the answer https://stackoverflow.com/a/66953015/ provided by the user 'Akshay Sehgal' ( https://stackoverflow.com/u/4755954/ ) at 'Stack Overflow' website. Thanks to these great users and Stackexchange community for their contributions.

Visit these links for original content and any more details, such as alternate solutions, latest updates/developments on topic, comments, revision history etc. For example, the original title of the Question was: How can I unnormalize MinMaxScaler?

Also, Content (except music) licensed under CC BY-SA https://meta.stackexchange.com/help/l...
The original Question post is licensed under the 'CC BY-SA 4.0' ( https://creativecommons.org/licenses/... ) license, and the original Answer post is licensed under the 'CC BY-SA 4.0' ( https://creativecommons.org/licenses/... ) license.

If anything seems off to you, please feel free to write me at vlogize [AT] gmail [DOT] com.
---
How to Unnormalize Your Data Using MinMaxScaler in Python

When working with machine learning models, it’s common to scale your features to improve performance or training speed. One popular method for scaling is the MinMaxScaler provided by the scikit-learn library in Python. However, after making predictions, you may need to convert those normalized values back to their original scale. This process is called "unnormalizing." In this guide, we will explore how to easily achieve this with MinMaxScaler.

The Problem at Hand

You might be facing a situation where you've already applied MinMaxScaler to your training data, as shown in the example code below:

[[See Video to Reveal this Text or Code Snippet]]

After you have scaled your data and predicted new outputs, you may want to convert your predictions back into the original value scale. This is where unnormalizing your data becomes necessary, and many users find themselves asking: How can I unnormalize MinMaxScaler data?

The Solution: Using the inverse_transform Method

To unnormalize your data after predictions, you can use the inverse_transform method that is part of the MinMaxScaler class. This method takes the scaled data and converts it back to the original form using the scaling parameters (min and max values) that were used during the fit process.

Here's how you can implement this in your code:

Step-by-Step Guide to Unnormalize Data

Fit the Scalers: First, ensure that you've properly fitted the MinMaxScaler to your training data. This is crucial as the scaler needs to know the original min and max values to revert the scaling later.

Transform the Test Data: Use the transform method for any data you want to predict, including test data or new unseen data.

Make Predictions: After training your model, you will predict the test values which will be normalized.

Inverse Transform Predicted Data: Finally, you will apply the inverse_transform method on the predicted values to obtain the original values.

Example Code

Here’s how you can achieve unnormalization in Python:

[[See Video to Reveal this Text or Code Snippet]]

Explanation of the Code

model.predict(X_test): This line invokes your machine learning model to make predictions on the unseen test data. These predictions will be in normalized form, as per the scaling applied earlier.

scaler_Y.inverse_transform(predicted_normalized): Here, you take the normalized predictions and convert them back to the original scale using the scaler fitted on your training y-values.

Conclusion

Unnormalizing data is an essential step in interpreting the results of your machine learning models where scaling was applied. With just a few lines of code using the inverse_transform method from the MinMaxScaler, you can easily revert predictions back to their original values. This allows you to present your results accurately and make informed decisions based on the original scale.

Happy Coding!