A learning curve visualizes model performance (training and validation scores) as a function of training set size. It is essential for diagnosing bias vs variance tradeoffs, determining whether collecting more data would improve model performance, and guiding model selection decisions. The plot typically shows two lines with shaded confidence bands representing variability across cross-validation folds.
""" anyplot.ai
learning-curve-basic: Model Learning Curve
Library: letsplot 4.9.0 | Python 3.13.13
Quality: 90/100 | Updated: 2026-05-10
"""
import os
import numpy as np
import pandas as pd
from lets_plot import *
from lets_plot import ggsave
LetsPlot.setup_html()
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
ELEVATED_BG = "#FFFDF6" if THEME == "light" else "#242420"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
# Okabe-Ito palette (first series always #009E73)
IMPRINT = ["#009E73", "#C475FD", "#4467A3", "#BD8233", "#AE3030", "#2ABCCD", "#954477"]
# Data - Simulate learning curve for a model showing slight overfitting pattern
np.random.seed(42)
# Training set sizes (10 different sizes)
train_sizes = np.array([50, 100, 200, 400, 600, 800, 1000, 1200, 1400, 1600])
# Simulate 5 cross-validation folds
n_folds = 5
n_sizes = len(train_sizes)
# Training scores: Start high, stay high (model fits training data well)
train_scores_mean = 0.99 - 0.15 * np.exp(-train_sizes / 200)
train_scores = np.zeros((n_folds, n_sizes))
for i in range(n_folds):
noise = np.random.randn(n_sizes) * 0.01
train_scores[i] = train_scores_mean + noise
# Validation scores: Start lower, improve with more data (learning effect)
# Show a gap with training that narrows as data increases
validation_scores_mean = 0.65 + 0.20 * (1 - np.exp(-train_sizes / 500))
validation_scores = np.zeros((n_folds, n_sizes))
for i in range(n_folds):
noise = np.random.randn(n_sizes) * 0.02
validation_scores[i] = validation_scores_mean + noise
# Calculate means and standard deviations
train_mean = np.mean(train_scores, axis=0)
train_std = np.std(train_scores, axis=0)
val_mean = np.mean(validation_scores, axis=0)
val_std = np.std(validation_scores, axis=0)
# Create DataFrames for plotting
df_train = pd.DataFrame(
{
"Training Set Size": train_sizes,
"Score": train_mean,
"Lower": train_mean - train_std,
"Upper": train_mean + train_std,
"Type": "Training Score",
}
)
df_val = pd.DataFrame(
{
"Training Set Size": train_sizes,
"Score": val_mean,
"Lower": val_mean - val_std,
"Upper": val_mean + val_std,
"Type": "Validation Score",
}
)
df = pd.concat([df_train, df_val], ignore_index=True)
# Plot
plot = (
ggplot(df, aes(x="Training Set Size", y="Score", color="Type", fill="Type"))
+ geom_ribbon(aes(ymin="Lower", ymax="Upper"), alpha=0.2, color="rgba(0,0,0,0)")
+ geom_line(size=2)
+ geom_point(size=4)
+ scale_color_manual(values=IMPRINT[:2])
+ scale_fill_manual(values=IMPRINT[:2])
+ scale_y_continuous(limits=[0.55, 1.02])
+ scale_x_continuous(limits=[0, 1700], breaks=list(range(0, 1800, 200)))
+ labs(
x="Training Set Size (samples)",
y="Accuracy Score (0-1)",
title="learning-curve-basic · letsplot · anyplot.ai",
color="",
fill="",
)
+ theme_minimal()
+ theme(
plot_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
panel_background=element_rect(fill=PAGE_BG),
panel_grid_major=element_line(color=INK_SOFT, size=0.3),
panel_grid_minor=element_blank(),
axis_title=element_text(color=INK, size=20),
axis_text=element_text(color=INK_SOFT, size=16),
axis_line=element_line(color=INK_SOFT),
plot_title=element_text(color=INK, size=24),
legend_background=element_rect(fill=ELEVATED_BG, color=INK_SOFT),
legend_text=element_text(color=INK_SOFT, size=16),
legend_title=element_text(color=INK),
legend_position="bottom",
)
+ ggsize(1600, 900)
)
# Save as PNG (scale 3x = 4800 x 2700 px) and HTML
ggsave(plot, f"plot-{THEME}.png", path=".", scale=3)
ggsave(plot, f"plot-{THEME}.html", path=".")
Part of Model Learning Curve on anyplot.ai.