Environment:Fastai Fastbook Sklearn Environment

Knowledge Sources	fastai/fastbook scikit-learn
Domains	Tabular, Machine_Learning
Last Updated	2026-02-09 17:00 GMT

Overview

Scikit-learn environment required for Random Forest models, decision tree visualization, and feature importance analysis in Ch9 Tabular Modeling.

Description

The Tabular Modeling chapter (Ch9) uses scikit-learn's `RandomForestRegressor` and `DecisionTreeRegressor` alongside fastai's tabular deep learning tools. The `utils.py` file also imports `sklearn.tree.export_graphviz` for decision tree visualization. Scikit-learn is listed in `requirements.txt` as `scikit_learn` (no version constraint).

Usage

Use this environment for the Tabular Modeling workflow, specifically for:

Random Forest training: `RandomForestRegressor` with custom hyperparameters
Decision tree visualization: `export_graphviz` rendered via graphviz
Feature importance analysis: Permutation importance and partial dependence
Model ensembling: Comparing RF predictions with neural network predictions

System Requirements

Category	Requirement	Notes
OS	Any (Linux, macOS, Windows)	No platform restrictions
Hardware	CPU only	Random forests do not use GPU
RAM	4GB+ recommended	For large tabular datasets with many trees

Dependencies

Python Packages

`scikit_learn` (any recent version)
`pandas` (for data manipulation)
`graphviz` (for decision tree visualization)
`scipy` (for `scipy.cluster.hierarchy` and `scipy.stats.spearmanr`)

Credentials

No credentials required.

Quick Install

pip install scikit-learn graphviz scipy

Code Evidence

Random Forest wrapper from `09_tabular.md:617-621`:

def rf(xs, y, n_estimators=40, max_samples=200_000,
       max_features=0.5, min_samples_leaf=5, **kwargs):
    return RandomForestRegressor(n_jobs=-1, n_estimators=n_estimators,
        max_samples=max_samples, max_features=max_features,
        min_samples_leaf=min_samples_leaf, oob_score=True).fit(xs, y)

Decision tree visualization from `utils.py:81-86`:

from sklearn.tree import export_graphviz

def draw_tree(t, df, size=10, ratio=0.6, precision=0, **kwargs):
    s=export_graphviz(t, out_file=None, feature_names=df.columns, filled=True, rounded=True,
                      special_characters=True, rotate=False, precision=precision, **kwargs)
    return graphviz.Source(re.sub('Tree {', f'Tree {{ size={size}; ratio={ratio}', s))

Scipy clustering from `utils.py:90-98`:

from scipy.cluster import hierarchy as hc

def cluster_columns(df, figsize=(10,6), font_size=12):
    corr = np.round(scipy.stats.spearmanr(df).correlation, 4)
    corr_condensed = hc.distance.squareform(1-corr)
    z = hc.linkage(corr_condensed, method='average')

Common Errors

Error Message	Cause	Solution
`ModuleNotFoundError: No module named 'sklearn'`	scikit-learn not installed	`pip install scikit-learn`
`ExecutableNotFoundError: graphviz not found`	System graphviz not installed	Install via package manager: `apt install graphviz` or `brew install graphviz`
`MemoryError` during RF training	Dataset too large for available RAM	Reduce `max_samples` parameter or use fewer trees

Compatibility Notes

n_jobs=-1: The Random Forest code uses `n_jobs=-1` to parallelize across all CPU cores. This works on all platforms.
Graphviz: Requires both the Python package (`pip install graphviz`) and the system binary (`apt install graphviz` on Linux).
No GPU: Scikit-learn models run exclusively on CPU. GPU is not used for random forests.

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment