Classifiers, Regressors & Clusters: The 3 Shapes of ML

Almost every ML problem is one of three shapes. Learn to recognise each — with a runnable example for all three.

AI & ML2026-07-219 min readscikit-learn · Python · ML

In this guide

  1. 1. Classification — which category?
  2. 2. Regression — how much?
  3. 3. Clustering — what groups exist?
  4. How to pick the right one

Machine learning looks sprawling until you notice almost everything is one of three problems: predict a category (classification), predict a number (regression), or find structure (clustering). Master these and you can frame most real problems.

1. Classification — which category?

python
from sklearn.datasets import load_iris
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split

X, y = load_iris(return_X_y=True)
Xtr, Xte, ytr, yte = train_test_split(X, y, test_size=0.2, random_state=0)

clf = RandomForestClassifier(n_estimators=100)
clf.fit(Xtr, ytr)
print("accuracy:", clf.score(Xte, yte))

Use it for spam detection, churn prediction, image labels, credit risk — anything with a finite set of labels.

2. Regression — how much?

python
from sklearn.datasets import load_diabetes
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error

X, y = load_diabetes(return_X_y=True)
Xtr, Xte, ytr, yte = train_test_split(X, y, test_size=0.2, random_state=0)

reg = LinearRegression().fit(Xtr, ytr)
pred = reg.predict(Xte)
print("MAE:", mean_absolute_error(yte, pred))

Use it for pricing, forecasting, sensor readings, energy use — any continuous quantity.

3. Clustering — what groups exist?

python
from sklearn.datasets import make_blobs
from sklearn.cluster import KMeans

X, _ = make_blobs(n_samples=300, centers=3, random_state=0)
km = KMeans(n_clusters=3, n_init=10).fit(X)
print("cluster labels:", km.labels_[:10])
print("inertia (lower is tighter):", km.inertia_)

Use it for customer segmentation, anomaly detection, topic grouping — finding structure when you have no labels.

How to pick the right one

python
# a quick checklist to ask yourself:
# 1. Do I have labels for past examples?
#      yes -> supervised (classification or regression)
#      no  -> unsupervised (clustering)
# 2. Is the thing I predict a category or a number?
#      category -> classification
#      number   -> regression
Always split your data, and hold out a test set the model never saw. A model that scores 99% on its own training data is usually just memorising.

Three shapes, one library, a lifetime of problems. The hard part is never the algorithm — it's framing the question and cleaning the data.

Keep reading

More from the lab

Python

Advanced Python Data Structures for Interviews & Code

collections, defaultdicts, heaps and deques — the structures that show up in every interview and codebase.

Python · Data Structures · 8 min AI & ML

Fine-Tuning a Large Language Model: A Realistic Guide

When fine-tuning is worth it, how LoRA works, and when RAG is the better answer.

LLM · Fine-tuning · Python · 9 min Data

PostgreSQL vs MySQL: Choosing a Database

The real differences that matter — integrity, JSON, extensions, scaling — and which to pick.

PostgreSQL · MySQL · SQL · 7 min
Your move

Want us to build this for you?

We turn tutorials like these into production systems every day. Tell us what you need.

Start a project →