How deeply does Statistics Every Programmer Needs cover feature set?

Shows column selection, label separation, reproducible splitting, and random feature consideration, but does not derive feature-selection criteria.

Feature set

As explained in Statistics Every Programmer Needs

A feature subset is a selected collection of predictor variables considered for a model operation such as a tree split.

A feature set is the selected collection of predictor variables supplied to a machine-learning model. In a classification workflow, it is separated from the target variable before training and evaluation. section 6.3.1

A practical feature-set workflow includes:

Select relevant predictor columns and assign them to X. section 6.3.1
Keep the target labels in y, separate from X. section 6.3.1
Split X and y into training and test sets to measure generalization and detect overfitting. section 6.3.1
In a random forest, evaluate a random subset of features at each tree split to increase model diversity. section 6.4.1

Football conversion model

A conversion classifier uses QUARTER, TO_GO, OFFENSIVE_TEAM_VENUE, SCORE_DIFFERENTIAL, and PLAY_TYPE as its feature set, while CONVERT supplies the labels.

How Statistics Every Programmer Needs covers it

Cover of Statistics Every Programmer Needs

liveBook manning.com

Read Statistics Every Programmer Needs on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Graph Algorithms for Data ScienceA feature set is a collection of graph relationships retained to calculate network features for predictive modeling.

Compare how 2 books explain feature set →

Related concepts

configurationconfigured by · A specification of settings, parameters, or component relationships that determines how a system, application, model, or process is structured or operates.
random forestuses · A random forest is an ensemble model that combines predictions from multiple decision trees built with randomized data samples, feature selections, or both.
deep learningused by · Deep learning is a machine-learning approach that uses neural networks with multiple computational layers to learn representations from data.
gradient boostingused by · Gradient boosting is an ensemble-learning technique that sequentially adds predictive models, typically weak learners, to reduce a loss by fitting its gradient.
feature selectioncontrasts with · Feature selection is the process of choosing a subset of input features for use in a modeling, analysis, or prediction task.

This explainer was generated by Manning's knowledge graph from the text of Statistics Every Programmer Needs. The section links point to the parts of the book it draws on.