How deeply does Math and Architectures of Deep Learning cover min-max normalization?

Defines the column-wise formula and shows its PyTorch implementation, stopping short of discussing edge cases or inverse scaling.

Min-max normalization

As explained in Math and Architectures of Deep Learning

A preprocessing method that rescales values into a fixed range, here 0–1.

Min-max normalization rescales feature values to a common 0–1 range. Aligning differently sized inputs, such as sales counts and discount percentages, can speed neural-network training and reduce the chance of getting stuck at a local minimum. section 8.5

For each feature column, subtract its minimum and divide by its maximum minus its minimum: X′ = (X − Xmin) / (Xmax − Xmin). The smallest observed value becomes 0 and the largest becomes 1. The PyTorch routine applies the same transformation to the target values.

Demand prediction data

A demand-prediction dataset combines previous-week sales, discount fraction, weeks to the next holidays, and units sold. Sales values are much larger than percentage or count features, making a shared 0–1 scale useful before training the neural network. section 8.5

How Math and Architectures of Deep Learning covers it

Cover of Math and Architectures of Deep Learning

liveBook manning.com

Read Math and Architectures of Deep Learning on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Grokking AI Algorithms, Second EditionMin-max scaling is a feature-scaling technique that linearly maps values between a specified minimum and maximum to a normalized range, commonly 0 to 1.
liveBook manning.com

Outlier Detection in PythonA feature-scaling transformation that linearly maps values to a specified interval, commonly from zero to one.
liveBook manning.com

The Complete Obsolete Guide to Generative AIMin-max scaling is a normalization method that linearly transforms values into a specified range, commonly from 0 to 1.

Compare how 4 books explain min-max normalization →

Related concepts

normalizationused by · Normalization is the process of transforming a value, representation, or data structure to satisfy a specified standard, scale, or structural constraint.
predict_probaused by · predict_proba is a classifier method that returns estimated probabilities for each possible class for supplied input samples.
feature scalingimplements · Feature scaling is the transformation of numerical features to comparable magnitudes, ranges, or distributions for computation or model fitting.
anomaly scoreuses · An anomaly score is a numerical measure of how anomalous an observation or event is relative to a reference or expected pattern.
scikit-learnpart of · scikit-learn is an open-source Python library providing tools for classical machine learning, including algorithms, preprocessing, model selection, and evaluation.
RobustScaleralternative to · RobustScaler is a scikit-learn transformer that scales features using robust statistics such as the median and interquartile range.

This explainer was generated by Manning's knowledge graph from the text of Math and Architectures of Deep Learning. The section links point to the parts of the book it draws on.