AI School
01Home02About
05For teams06Contact
Get started→
Machine learning✦LLMs✦Agents✦Evals✦RAG✦Prompting✦Product✦Career✦Community✦
Machine learning✦LLMs✦Agents✦Evals✦RAG✦Prompting✦Product✦Career✦Community✦

[SIGNAL] The weekly brief

Weekly AI notes worth saving.

One short email. Curated frameworks, new tools, and the most useful papers — minus the hype.

Free forever2 min readUnsubscribe anytime
>_
AI School

The launchpad for tomorrow’s AI pioneers. Master artificial intelligence with world-class programs, mentors, and a community that moves you forward.

Learners5,000+
Mentors30+
Countries20+
01

Learn

  • Programs
  • AI PM Certificate
  • Career Guidance
  • Blog
  • About Us
  • Testimonials
02

Free Tools

  • AI Learning Path
  • Career Quiz
  • Algorithm Selector
  • AI Glossary
  • LLMs List
03

Company

  • For teams
  • Contact
  • LLM Email Course
04

Legal

  • Privacy Policy
  • Terms of Service

© 2026 Artificial Intelligence School — All systems operational

Read stories ↗Talk to us ↗
[TOOL 03] — ML algorithm selector9 problem classes indexed

ML algorithm
selector.

Not sure which algorithm to use? Tell us your problem type and we’ll recommend the best machine learning algorithms for your use case.

[01]Select your problem type

What’s your problem type?

Select the type of machine learning problem you’re trying to solve

AWAITING INPUT

STEP 02 / 03 — CONSTRAIN THE SHORTLIST (OPTIONAL)

How much data do you have?

Do decisions need to be explainable?

STEP 03 / 03 — RUN RECOMMENDATION [○]

[03]The method

How practitioners actually choose.

The selector above gives you a shortlist. This is the reasoning that turns a shortlist into a decision - the six steps that come up in every serious model review.

STEP / 01

Name the prediction, not the technique

Write the sentence "given X, predict Y, measured by Z" before touching a library. Whether Y is a category, a quantity, a ranking, or a structure decides 80% of the shortlist - and Z (the metric the business actually feels) decides the rest. Teams that skip this step tune the wrong loss beautifully.

STEP / 02

Let the data pick the family

Tabular rows with mixed types → gradient-boosted trees first, linear models for the baseline. Images, audio, raw text → pretrained deep models, fine-tuned, never from scratch. A single time series → statistical models before deep ones. Under ~1,000 samples → simpler beats cleverer, almost without exception.

STEP / 03

Price your constraints before accuracy

Interpretability requirements (regulated decisions), latency budgets (real-time serving), training cadence (nightly retrains), and memory ceilings (on-device) eliminate more candidates than any benchmark. An algorithm you cannot deploy, explain, or retrain is not an option - it is a demo.

STEP / 04

Establish the dumb baseline

Majority class. Yesterday’s value. The mean. A 10-line logistic regression. Every serious team measures lift over a baseline a child could describe, because half of all "ML wins" evaporate against one. Your first real model has one job: beat the dumb baseline honestly.

STEP / 05

Iterate on data before models

Once a sensible model is in place, an hour spent on features, label quality, and leakage checks buys more accuracy than an hour of hyperparameter search - a ratio practitioners put at 10:1. Swap algorithms when the current one plateaus with clean data, not before.

STEP / 06

Validate the way you will be judged

Random splits lie whenever time, users, or groups structure your data. Split by time for forecasting, by user for personalisation, by hospital for medical models. The gap between naive CV and honest CV is where production surprises live.

[04]Reference matrix

Every algorithm, one table.

21 ALGORITHMS INDEXED

AlgorithmFamilyInterpretabilityTraining speedData neededReach for it when
Linear RegressionRegressionHighFastSmallSimple prediction problems
Logistic RegressionClassificationHighFastSmallBinary classification
Decision TreeClassification/RegressionHighFastSmallInterpretable models
Random ForestClassification/RegressionMediumMediumMediumGeneral-purpose classification
Gradient Boosting (XGBoost/LightGBM)Classification/RegressionMediumMediumMediumKaggle competitions
Support Vector Machine (SVM)Classification/RegressionLowSlowMediumHigh-dimensional data
K-Nearest Neighbors (KNN)Classification/RegressionHighFastSmallRecommendation systems
Naive BayesClassificationHighFastSmallText classification
K-Means ClusteringClusteringHighFastMediumCustomer segmentation
DBSCANClusteringMediumMediumMediumSpatial data
Hierarchical ClusteringClusteringHighSlowSmallTaxonomy creation
Principal Component Analysis (PCA)Dimensionality ReductionMediumFastMediumData visualization
t-SNEDimensionality ReductionLowSlowMedium2D/3D visualization
Neural Network (MLP)Deep LearningLowSlowLargeComplex patterns
Convolutional Neural Network (CNN)Deep LearningLowSlowLargeImage classification
RNN/LSTMDeep LearningLowSlowLargeTime series
TransformerDeep LearningLowSlowLargeNLP tasks
Isolation ForestAnomaly DetectionMediumFastMediumFraud detection
Collaborative FilteringRecommendationMediumMediumLargeE-commerce
ARIMATime SeriesHighFastMediumEconomic forecasting
ProphetTime SeriesHighFastMediumBusiness forecasting

[06]Data profiler

Or just show us the data.

Drop a CSV and get a column-by-column profile plus a recommendation. Everything runs in your browser - the file never leaves your machine.

[05]Honest answers

The questions everyone asks.

Q.01 — Q.07

Q.01Which machine learning algorithm should a beginner start with?+

Linear or logistic regression, seriously. They train in seconds, their coefficients teach you what the data is doing, and they set an honest baseline. Then learn decision trees for non-linear intuition, then random forests and gradient boosting - the pair that wins most real tabular problems.

Q.02When should I use deep learning instead of classical ML?+

When the input is unstructured - images, audio, video, free text - or when you can fine-tune a large pretrained model. On ordinary tabular business data, gradient-boosted trees match or beat neural networks in most published benchmarks while training faster and needing less tuning.

Q.03What is the best algorithm for small datasets?+

Under a few thousand rows: regularised linear/logistic regression, naive Bayes for text, or a shallow tree. Strong regularisation plus cross-validation matters more than the algorithm. Deep networks and unconstrained boosting will mostly memorise noise at this scale.

Q.04Random forest vs gradient boosting - which one?+

Boosting (XGBoost/LightGBM/CatBoost) is usually a few points more accurate when tuned with early stopping; forests are nearly tuning-free, parallel, and degrade more gracefully. Rule of thumb: forest for a robust v1 you barely babysit, boosting when accuracy is worth tuning care.

Q.05How much data do I need for each type of algorithm?+

Rough orders of magnitude: linear models and naive Bayes work from hundreds of rows; trees and forests want thousands; gradient boosting shines from thousands to millions; training deep networks from scratch wants tens of thousands to millions - though fine-tuning a pretrained model can need only hundreds.

Q.06Do I need to normalise or scale my features?+

For distance- and gradient-based methods - kNN, SVM, k-means, PCA, neural networks, regularised regression - yes, always. Tree-based models (decision trees, random forests, gradient boosting) are scale-invariant: they split on thresholds, so scaling changes nothing.

Q.07What actually matters more than the algorithm choice?+

Label quality, leakage prevention, honest validation splits, and features that encode domain knowledge. Practitioners consistently report bigger gains from an hour of data work than an hour of model swapping. The algorithm decides your ceiling; the data decides whether you ever reach it.