Machine Learning Terms Every Beginner Should Understand

Core Terms — Plain English:
• Features → The input columns
• Label → The output we predict
• Dataset → Rows of features (+ labels)
• Supervised Learning → Learning from labels
• Unsupervised Learning → Finding patterns without labels
• Classification → Predicting a category
• Regression → Predicting a number

Where You'll See Them:
☐ In dataset descriptions
☐ In model documentation
☐ In paper abstracts
☐ In interview questions
☐ In ML course notebooks
☐ In feature engineering discussions

Training Terms — Plain English:
• Algorithm → The method ("how to learn")
• Model → The algorithm + learned weights
• Training → Fitting the model to data
• Hyperparameters → Settings you choose
• Parameters → Settings the model learns
• Training Set → Where the model learns
• Validation Set → Where you tune
• Test Set → Where you judge

Typical ML Workflow:
☐ Split data → train / validation / test
☐ Choose an algorithm
☐ Set hyperparameters
☐ Fit the model on training data
☐ Evaluate on validation set
☐ Tune hyperparameters
☐ Final evaluation on test set
☐ Deploy the model

Evaluation Terms — Plain English:
• Overfitting → Great on train, bad on test
• Underfitting → Bad on both train and test
• Accuracy → Overall correct rate
• Precision → How often positives are correct
• Recall → How many real positives we caught
• F1 Score → Balance of precision and recall
• Confusion Matrix → Full detail of errors
• Cross-Validation → Test on multiple splits

Precision vs Recall Quick Guide:
Precision matters when false positives are costly
(e.g., spam filters — don't block real email)

Recall matters when false negatives are costly
(e.g., cancer screening — don't miss a case)

F1 combines both — use when you need balance.

Signals:
□ Train accuracy high + test low → overfitting
□ Both low → underfitting
□ Class imbalance → use precision/recall/F1

Term Mastery Tracker:
☐ Features — can identify in a dataset
☐ Label — can identify in a dataset
☐ Dataset — can explain training vs test
☐ Supervised Learning — can give example
☐ Unsupervised Learning — can give example
☐ Classification — can give example
☐ Regression — can give example
☐ Algorithm — can name 2 examples
☐ Model — can explain learned parameters
☐ Hyperparameters — can name 2 examples
☐ Parameters — can distinguish from hyperparams
☐ Training Set — can describe its purpose
☐ Validation Set — can describe its purpose
☐ Test Set — can describe its purpose
☐ Overfitting — can spot in a metric curve
☐ Underfitting — can spot in a metric curve
☐ Accuracy — can explain its limits
☐ Precision — can explain when to use it
☐ Recall — can explain when to use it
☐ F1 Score — can explain the trade-off
☐ Confusion Matrix — can interpret one

Weekly Review — Term Check:
Week: _____

Terms I fully understand:
1. 
2. 
3. 

Terms I'm still shaky on:
1. 
2. 
3. 

Terms I used in code this week:
1. 
2. 
3. 

Next week's focus terms:
1. 
2. 
3. 

Concept Map — How Terms Connect:
Dataset (features + label)
  ↓
Split → Train / Validation / Test
  ↓
Choose Algorithm + set Hyperparameters
  ↓
Train → Model learns Parameters
  ↓
Evaluate on Validation Set
  ↓
Tune Hyperparameters
  ↓
Final evaluation on Test Set
  ↓
Check overfitting / underfitting
  ↓
Report Accuracy, Precision, Recall, F1

If you can draw this, you understand the terms.