Logistic regression
Data
Data
Dataset
Auto MPG
Iris
Palmer Penguins
Diabetes
Linear
Quadratic
Linear with outliers
Load CSV
Input feature
mpg
cylinders
displacement
horsepower
weight
acceleration
model_year
Target feature
mpg
cylinders
displacement
horsepower
weight
acceleration
model_year
Standardize features
Class 1 when mpg >
Source
Observable notebook
· R. Quinlan · UCI · CC BY 4.0
Model
Model
x² coefficient
0
Regularization
None
L1
L2
Optimization
Optimization
Optimizer
Gradient descent
Momentum
Nesterov
Adagrad
RMSprop
Adam
BFGS
Newton
Learning rate
Batches
Full batch
Mini-batch SGD
Start weight
Start bias
View
View
Landscape
Surface
2D
3D
Color scale
log(1 + loss)
Loss
Contours
Zoom
Plot settings…
Data plot
Residual lines
Density contours
Loss over (w, b)
log(1 + loss)
Fit
Loading...
bias: -4 … 4
weight: -8 … 8
horsepower → mpg
0 observations
p(y = 1 | x) = σ(1.000x + 0.000)
0
0.5
1
−3
−2
−1
0
1
2
3
Gray: starting fit · Rose: current fit · Point color: loss
Run
Animation speed
Smooth
Instant
steps/s
Step
0
Loss
—