Data balancing

As explained in Deep Learning with PyTorch, Second Edition

Adjusting class proportions in a dataset to reduce skew and make learning less biased toward a dominant class.

Data balancing adjusts class proportions so a classifier does not favor a dominant class. This matters when heavily skewed negative samples let a model achieve high accuracy by predicting everything as negative. section 14.0

In the lung nodule classification workflow, balancing is treated as a training improvement alongside augmentation. The changes target the dataset implementation and are evaluated through performance metrics. section 14.1

Figure 14.1 Our end-to-end lung cancer detection project, with a focus on this chapter’s topic: step 3, classification.

Places balancing within the classification step of the end-to-end lung cancer detection project. section 14.0

How Deep Learning with PyTorch, Second Edition covers it

How deep it goes

Positions balancing as a LunaDataset change to improve training, but does not detail its procedure.

Examples

Uses lung nodule classification with heavily negative-skewed samples as its motivating example.

What it assumes

Assumes a working classification-training workflow and motivates balancing through misleading overall accuracy.

What you can do after

Apply it by modifying the dataset and checking whether performance metrics improve during training.

How it connects

Connects balancing with classification, augmentation, precision, recall, and F1 score.

liveBookmanning.com

Read Deep Learning with PyTorch, Second Edition on liveBook Get the book at manning.com

Related concepts

• upsamplinguses · Upsampling is a process that increases the number, density, or spatial or temporal resolution of samples or values in a dataset or representation by adding, duplicating, or estimating them.
• datasetimplemented by · A dataset is an organized collection of related data items, records, or observations treated as a unit for storage, processing, analysis, or computation.
• DataLoaderalternative to · PyTorch DataLoader is an iterable utility that retrieves dataset samples, optionally shuffles and transforms them, and groups them into batches.

This explainer was generated by Manning's knowledge graph from the text of Deep Learning with PyTorch, Second Edition. The section links point to the parts of the book it draws on.