Feature Importance Calculator

Estimate how many features to keep, overfitting risk, and expected variance retention based on dataset size, model type, and correlation threshold.

Inputs

Total Features

Dataset Size (rows)

Model Type

Target Variance Explained (%)

Correlation Threshold

Tab to move between fields

Results

Recommended Features to Keep

Est. Variance Retained

100%

Features to Remove0

Samples per Feature20

Min Samples Needed750

Overfitting Risk8%

How to Use This Calculator

Enter Total Features (number of columns in your dataset).
Enter Dataset Size (number of rows/samples).
Select your Model Type — each type has different feature capacity.
Review Recommended Features to Keep and Overfitting Risk.
Check if your dataset meets the Minimum Samples Needed.
Use the chart to see the keep vs. remove split.

How the result changes with Total Features

Total Features	Recommended Features to Keep	Est. Variance Retained
10,002	100	2.1%
35,001	100	0.6%
65,001	100	0.3%
90,000	100	0.2%

What each input means

Total Features: Total number of features in the dataset
Dataset Size (rows): Number of samples/rows in the dataset
Model Type: Model type affects feature selection heuristics and data requirements
Target Variance Explained (%): Desired percentage of variance to retain after feature selection
Correlation Threshold: Features with correlation above this are considered redundant

How this is calculated

Worked example, using the default values

Assess Data-to-Feature Ratio
effectiveFeatureRatio = datasetSize ÷ totalFeatures
1000 ÷ 50 = 20 = 20 samples per feature (Very Low overfitting risk)
Calculate Recommended Features
recommendedFeatures = min(totalFeatures, datasetSize ÷ 10)
min(50, 1000 ÷ 10) = min(50, 100) = 50 features (Linear/Regularized)
Estimate Variance Retained
V(k) = [1 − (1 − k/n)^α] × 100
[1 − (1 − 50/50)^1.8] × 100 = 100% variance retained with 50 of 50 features
Calculate Minimum Samples Needed
minSamples = totalFeatures × samplesPerFeature (15× for Linear/Regularized)
50 × 15 = 750 samples needed (you have 1000 ✓)

Engine last updated August 5, 2026. Checked against 2 independently-derived tests — how we verify calculators.

The questions that sit next to this one — chosen by subject, including calculators filed under a different category.

MLOps & AI Costing

Training Data Size Calculator

Estimate minimum dataset size needed for machine learning models based on features, complexity, and accuracy targets.

MLOps & AI Costing

Hyperparameter Grid Search Calculator

Calculate total experiments, time, and GPU cost for hyperparameter grid search with cross-validation.

Machine Learning & AI

Information Gain Calculator

Calculate information gain, gain ratio, Gini gain, and feature importance for decision trees.

Machine Learning & AI

Overfitting Calculator

Calculate overfitting metrics, generalization gap, bias-variance trade-off, and model complexity analysis.

Data Science

Dimensionality Reduction Calculator

Determine how many principal components to retain based on eigenvalues and target explained variance. Includes scree plot data, Kaiser criterion, and cumulative variance analysis.