cloudsandai
Multiple Linear Regression: An Interactive Learning Lab
Explore how multiple features combine in a linear model, see how feature scaling affects training, and fit predictions with gradient descent.
Section A
Explore the dataset
Estimate a home's price in $10,000 units using four features. Incomplete rows are ignored until all values are entered.
| Area (100 sq ft) | Bedrooms | Age (years) | Commute (minutes) | Price ($10k) | Action |
|---|---|---|---|---|---|
10 complete observations
Section C
Fit and evaluate
Ready for training. Step 0
MSE
4131.100
R squared
-11.742
Valid rows
10
Prediction breakdown
Feature contributions
Observation 1: each term shows its contribution to this prediction.
Area (100 sq ft)8.00 (z -1.14)0.00
Bedrooms2.00 (z -1.17)0.00
Age (years)12.00 (z -0.09)0.00
Commute (minutes)18.00 (z -0.07)0.00
Bias0.00
Predicted price0.00 ($10k)
Section D
Predicted versus actual
Points closer to the diagonal represent more accurate predictions.
Observations Perfect prediction
| Observation | Actual ($10k) | Predicted ($10k) | Residual |
|---|---|---|---|
| 1 | 43.00 | 0.00 | -43.00 |
| 2 | 55.00 | 0.00 | -55.00 |
| 3 | 37.00 | 0.00 | -37.00 |
| 4 | 65.00 | 0.00 | -65.00 |
| 5 | 74.00 | 0.00 | -74.00 |
| 6 | 45.00 | 0.00 | -45.00 |
| 7 | 86.00 | 0.00 | -86.00 |
| 8 | 57.00 | 0.00 | -57.00 |
| 9 | 96.00 | 0.00 | -96.00 |
| 10 | 59.00 | 0.00 | -59.00 |
Training history
Loss versus step
Starting MSE: 4131.100 | Current MSE: 4131.100