📘 Lesson · Lesson 89
Overfitting & Underfitting
Overfitting & Underfitting
Model Fitting Problems
💡 At a Glance
Overfitting = model memorizes training data but fails on new data. Underfitting = model is too simple to learn the pattern.
Comparison
| Problem | Training | New Data | Fix |
|---|---|---|---|
| Overfitting | very good | poor | more data, regularization, simpler model |
| Underfitting | poor | poor | more features, complex model, train longer |
Summary
- Overfitting: great on training, bad on new data — simplify or add data.
- Underfitting: bad everywhere — use a richer model or more features.
Model Fitting समस्याएं
💡 एक नज़र में
Overfitting = model training data रट लेता है पर नए data पर fail। Underfitting = model इतना simple कि pattern सीख ही नहीं पाता।
तुलना
| Problem | Training | New Data | Fix |
|---|---|---|---|
| Overfitting | बहुत अच्छा | खराब | ज़्यादा data, regularization, simpler model |
| Underfitting | खराब | खराब | ज़्यादा features, complex model, ज़्यादा train |
सारांश
- Overfitting: training पर बढ़िया, नए data पर खराब — simplify या data जोड़ें।
- Underfitting: हर जगह खराब — richer model या ज़्यादा features।