TorchGMM Plotting Function Comprehensive Demo¶

This notebook demonstrates all the features and parameters of the plot_gmm function in the TorchGMM library. We'll go through each parameter systematically to show how they affect the visualization.

Overview of Parameters:¶

The plot_gmm function provides extensive control over GMM visualization through several parameter categories:

Core Data Parameters

• X: Input data points to plot (array-like, shape (n_samples, 2))
• gmm: Fitted GMM object with predict(), means_, covariances_, etc. (optional)

Data Point Styling

• show_points (bool, default=True): Whether to display the data points
• point_size (float, default=5): Size of individual data points
• point_alpha (float, default=0.5): Transparency of points (0=invisible, 1=opaque)
• point_color (str/array, default='auto'): Point coloring strategy:
  • 'auto': Smart coloring based on available cluster info
  • 'black': All points in black
  • Colormap name: Use matplotlib colormap (e.g., 'viridis', 'plasma')
  • Single color: Use one color for all points (e.g., 'red', '#FF0000')
  • Array-like: Custom colors/values for each point

Cluster Visualization

• color_by_cluster (bool, default=False): Enable cluster-based point coloring
• true_labels (array-like, optional): Ground truth cluster labels for comparison
• match_labels_to_true (bool, default=False): Remap predicted labels to match true labels using Hungarian algorithm
• cluster_colors (str/list, default='turbo'): Cluster color specification:
  • Matplotlib colormap name (e.g., 'turbo', 'viridis')
  • Single color for all clusters
  • List of specific colors (e.g., ['red', 'blue', 'green'])
• show_incorrect_predictions (bool, default=False): Highlight correct (green) vs incorrect (red) predictions

Continuous Coloring

• log_probs (array-like, optional): Continuous values for color mapping (e.g., log-probabilities)
• colormap (str, default='viridis'): Matplotlib colormap for continuous values
• colorbar_label (str, default='Log Probability'): Label for the colorbar

Component Ellipses

• show_ellipses (bool, default=True): Display confidence ellipses for GMM components
• ellipse_std_devs (list, default=[1, 2, 3]): Standard deviation levels for ellipse boundaries
• ellipse_alpha (float, default=0.5): Fill transparency of ellipses
• ellipse_colors (str/list, default='auto'): Ellipse color specification (same options as cluster_colors)
• ellipse_fill (bool, default=True): Whether ellipses should be filled or outline-only
• ellipse_line_style (str, default='dotted'): Line style ('solid', 'dashed', 'dotted')
• ellipse_line_width (float, default=2): Thickness of ellipse boundaries
• ellipse_line_color (str, default='black'): Color of ellipse boundary lines
• ellipse_line_alpha (float, default=0.5): Transparency of ellipse boundaries

Component Centers/Means

• show_means (bool, default=True): Display component center points
• mean_marker (str, default='h'): Marker style for means ('x', 'o', '*', 'h', etc.)
• mean_size (float, default=25): Size of mean markers
• mean_color (str, default='black'): Color of mean markers

Initial Means (if available)

• show_initial_means (bool, default=False): Show starting positions from initialization
• initial_mean_marker (str, default='H'): Marker style for initial means
• initial_mean_size (float, default=25): Size of initial mean markers
• initial_mean_color (str, default='red'): Color of initial mean markers

Weight-Based Scaling

• scale_alpha_by_weight (bool, default=False): Scale ellipse transparency by component weight
• scale_size_by_weight (bool, default=False): Scale marker size by component weight

Plot Styling & Layout

• ax (matplotlib.Axes, optional): Axes object to plot on (creates new if None)
• title (str, default='GMM Visualization'): Plot title
• xlabel (str, default='Feature 1'): X-axis label
• ylabel (str, default='Feature 2'): Y-axis label
• legend (bool, default=True): Whether to show legend
• legend_labels (list, optional): Custom labels for legend entries

In [1]:

import torch
import numpy as np
import matplotlib.pyplot as plt
import os

os.chdir('../')

import tgmm

# reload the module to reflect recent changes
import importlib
importlib.reload(tgmm)

from tgmm import GaussianMixture, GMMInitializer, dynamic_figsize, plot_gmm

device = 'cuda' if torch.cuda.is_available() else 'cpu'
random_state = 42
np.random.seed(random_state)
torch.manual_seed(random_state)

if device == 'cuda':
    torch.cuda.manual_seed(random_state)
    print('CUDA version:', torch.version.cuda)
    print('Device:', torch.cuda.get_device_name(0))
else:
    print('Using CPU')

import torch import numpy as np import matplotlib.pyplot as plt import os os.chdir('../') import tgmm # reload the module to reflect recent changes import importlib importlib.reload(tgmm) from tgmm import GaussianMixture, GMMInitializer, dynamic_figsize, plot_gmm device = 'cuda' if torch.cuda.is_available() else 'cpu' random_state = 42 np.random.seed(random_state) torch.manual_seed(random_state) if device == 'cuda': torch.cuda.manual_seed(random_state) print('CUDA version:', torch.version.cuda) print('Device:', torch.cuda.get_device_name(0)) else: print('Using CPU')

CUDA version: 12.4
Device: NVIDIA GeForce RTX 4060 Laptop GPU

/home/asp/miniforge3/lib/python3.12/site-packages/torch/cuda/__init__.py:654: UserWarning: Can't initialize NVML
  warnings.warn("Can't initialize NVML")

In [2]:

n_samples = [800, 200, 1000, 1000]
centers = [np.array([0, 2]),
           np.array([2, -2]),
           np.array([0, 0]),
           np.array([2, 2])]
covs = [
    1.0 * np.eye(2),                    # spherical covariance
    0.5 * np.eye(2),                    # spherical covariance, fewer points
    np.array([[2, 0], [0, 0.5]]),       # diagonal covariance
    np.array([[0.2, 0.5], [0.5, 2]])    # full covariance
]

components = []
for n, center, cov in zip(n_samples, centers, covs):
    samples = np.dot(np.random.randn(n, 2), cov) + center
    components.append(samples)

X = np.vstack(components)
labels = np.concatenate([i * np.ones(n) for i, n in enumerate(n_samples)])
legend_labels = [f'Component {i+1}' for i in range(len(n_samples))]

n_features = X.shape[1]
n_components = len(n_samples)

# Convert to tensor (if needed for further processing)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
X_tensor = torch.tensor(X, dtype=torch.float32, device=device)

n_components = 4

# Initialize the GMM
gmm = GaussianMixture(
    n_components=n_components,
    covariance_type='full',
    tol=1e-6,
    reg_covar=1e-10,
    max_iter=1000,
    init_params='random',
    cov_init_method='eye',
    weights_init=None,
    means_init=None,
    covariances_init=None,
    n_init = 5,
    random_state=None,
    warm_start=False,
    verbose=False,
    verbose_interval=100,
    device='cpu',
)
# Fit the GMM
gmm.fit(X_tensor)

# Get predictions
y_pred = gmm.predict(X_tensor)

# Compute per-sample log-likelihoods
log_probs = gmm.score_samples(X_tensor)

# Get probabilities for each sample for each component
probs = gmm.predict_proba(X_tensor)
probs = probs.detach()

# Generate new samples
n_samples_to_generate = len(X)
gmm_samples, gmm_labels = gmm.sample(n_samples_to_generate)
gmm_samples = gmm_samples.detach().cpu().numpy()
gmm_labels = gmm_labels.detach().cpu().numpy()

# Compute probabilities for each generated sample
generated_probs = gmm.predict_proba(torch.tensor(gmm_samples, dtype=torch.float32).to(device))
generated_probs = generated_probs.detach().cpu().numpy()

n_samples = [800, 200, 1000, 1000] centers = [np.array([0, 2]), np.array([2, -2]), np.array([0, 0]), np.array([2, 2])] covs = [ 1.0 * np.eye(2), # spherical covariance 0.5 * np.eye(2), # spherical covariance, fewer points np.array([[2, 0], [0, 0.5]]), # diagonal covariance np.array([[0.2, 0.5], [0.5, 2]]) # full covariance ] components = [] for n, center, cov in zip(n_samples, centers, covs): samples = np.dot(np.random.randn(n, 2), cov) + center components.append(samples) X = np.vstack(components) labels = np.concatenate([i * np.ones(n) for i, n in enumerate(n_samples)]) legend_labels = [f'Component {i+1}' for i in range(len(n_samples))] n_features = X.shape[1] n_components = len(n_samples) # Convert to tensor (if needed for further processing) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") X_tensor = torch.tensor(X, dtype=torch.float32, device=device) n_components = 4 # Initialize the GMM gmm = GaussianMixture( n_components=n_components, covariance_type='full', tol=1e-6, reg_covar=1e-10, max_iter=1000, init_params='random', cov_init_method='eye', weights_init=None, means_init=None, covariances_init=None, n_init = 5, random_state=None, warm_start=False, verbose=False, verbose_interval=100, device='cpu', ) # Fit the GMM gmm.fit(X_tensor) # Get predictions y_pred = gmm.predict(X_tensor) # Compute per-sample log-likelihoods log_probs = gmm.score_samples(X_tensor) # Get probabilities for each sample for each component probs = gmm.predict_proba(X_tensor) probs = probs.detach() # Generate new samples n_samples_to_generate = len(X) gmm_samples, gmm_labels = gmm.sample(n_samples_to_generate) gmm_samples = gmm_samples.detach().cpu().numpy() gmm_labels = gmm_labels.detach().cpu().numpy() # Compute probabilities for each generated sample generated_probs = gmm.predict_proba(torch.tensor(gmm_samples, dtype=torch.float32).to(device)) generated_probs = generated_probs.detach().cpu().numpy()

Default Parameters¶

In [3]:

nrows, ncols = 1, 1
figsize = dynamic_figsize(nrows, ncols)
fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize)

plot_gmm(
    X,
    gmm=None,

    # Data point styling
    show_points=True,
    point_size=5,
    point_alpha=0.5,
    point_color='auto',  # 'auto', 'black', array-like, or colormap name
    
    # Cluster visualization
    color_by_cluster=False,
    true_labels=None,
    match_labels_to_true=False,
    cluster_colors='turbo',  # Can be colormap name, single color, or list of colors
    show_incorrect_predictions=False,  # replaces 'outliers' mode
    
    # Continuous coloring (replaces 'continuous' mode)
    log_probs=None,
    colormap='viridis',
    colorbar_label='Log Probability',
    
    # Component ellipses
    show_ellipses=True,
    ellipse_std_devs=[1, 2, 3],  # List of standard deviations to show
    ellipse_alpha=0.5,
    ellipse_colors='auto',  # 'auto' uses same as clusters
    ellipse_fill=True,
    ellipse_line_style='dotted',
    ellipse_line_width=2,
    ellipse_line_color='black',
    ellipse_line_alpha=0.5,
    
    # Component centers/means
    show_means=True,
    mean_marker='h',
    mean_size=25,
    mean_color='black',
    
    # Initial means (if provided)
    show_initial_means=False,
    initial_mean_marker='H',
    initial_mean_size=25,
    initial_mean_color='red',
    
    # Weight visualization
    scale_alpha_by_weight=False,
    scale_size_by_weight=False,
    
    # Plot styling
    ax=ax,
    title='GMM Visualization',
    xlabel='Feature 1',
    ylabel='Feature 2',
    legend=True,
    legend_labels=None,
)

plt.tight_layout() 
plt.show()

nrows, ncols = 1, 1 figsize = dynamic_figsize(nrows, ncols) fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize) plot_gmm( X, gmm=None, # Data point styling show_points=True, point_size=5, point_alpha=0.5, point_color='auto', # 'auto', 'black', array-like, or colormap name # Cluster visualization color_by_cluster=False, true_labels=None, match_labels_to_true=False, cluster_colors='turbo', # Can be colormap name, single color, or list of colors show_incorrect_predictions=False, # replaces 'outliers' mode # Continuous coloring (replaces 'continuous' mode) log_probs=None, colormap='viridis', colorbar_label='Log Probability', # Component ellipses show_ellipses=True, ellipse_std_devs=[1, 2, 3], # List of standard deviations to show ellipse_alpha=0.5, ellipse_colors='auto', # 'auto' uses same as clusters ellipse_fill=True, ellipse_line_style='dotted', ellipse_line_width=2, ellipse_line_color='black', ellipse_line_alpha=0.5, # Component centers/means show_means=True, mean_marker='h', mean_size=25, mean_color='black', # Initial means (if provided) show_initial_means=False, initial_mean_marker='H', initial_mean_size=25, initial_mean_color='red', # Weight visualization scale_alpha_by_weight=False, scale_size_by_weight=False, # Plot styling ax=ax, title='GMM Visualization', xlabel='Feature 1', ylabel='Feature 2', legend=True, legend_labels=None, ) plt.tight_layout() plt.show()

Simple Data Points Without GMM¶

In [4]:

nrows, ncols = 1, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Different Point Styling Options without True Labels')

# Different point colors
point_color_options = [
    ('auto', 'Auto (black by default)', {}),
    ('red', 'Single color (red)', {}),
]

for i, (color_option, title, extra_params) in enumerate(point_color_options):
    plot_gmm(
        X,
        gmm=None,
        point_color=color_option,
        point_size=8,
        point_alpha=0.7,
        title=title,
        show_ellipses=False,
        show_means=False,
        ax=axes[i],
        **extra_params
    )

plt.tight_layout()
plt.show()

nrows, ncols = 1, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Different Point Styling Options without True Labels') # Different point colors point_color_options = [ ('auto', 'Auto (black by default)', {}), ('red', 'Single color (red)', {}), ] for i, (color_option, title, extra_params) in enumerate(point_color_options): plot_gmm( X, gmm=None, point_color=color_option, point_size=8, point_alpha=0.7, title=title, show_ellipses=False, show_means=False, ax=axes[i], **extra_params ) plt.tight_layout() plt.show()

Cluster Coloring with True Labels¶

In [5]:

nrows, ncols = 2, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Cluster Visualization with True Labels')

cluster_options = [
    (True, 'black', 'Cluster coloring (black)'),
    (True, 'turbo', 'Cluster coloring (turbo)'),
    (True, 'viridis', 'Cluster coloring (viridis)'),
    (True, ['red', 'blue', 'green', 'orange'], 'Custom cluster colors'),
]

for i, (color_by_cluster, cluster_colors, title) in enumerate(cluster_options):
    plot_gmm(
        X,
        gmm=None,
        color_by_cluster=color_by_cluster,
        true_labels=labels,
        cluster_colors=cluster_colors,
        title=title,
        show_ellipses=False,
        show_means=False,
        ax=axes[i]
    )

plt.tight_layout()
plt.show()

nrows, ncols = 2, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Cluster Visualization with True Labels') cluster_options = [ (True, 'black', 'Cluster coloring (black)'), (True, 'turbo', 'Cluster coloring (turbo)'), (True, 'viridis', 'Cluster coloring (viridis)'), (True, ['red', 'blue', 'green', 'orange'], 'Custom cluster colors'), ] for i, (color_by_cluster, cluster_colors, title) in enumerate(cluster_options): plot_gmm( X, gmm=None, color_by_cluster=color_by_cluster, true_labels=labels, cluster_colors=cluster_colors, title=title, show_ellipses=False, show_means=False, ax=axes[i] ) plt.tight_layout() plt.show()

Confidence Ellipses¶

In [6]:

nrows, ncols = 2, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('GMM Visualization with Different Ellipse Options')

# Basic GMM with ellipses
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    title='GMM with Ellipses (Default)',
    ax=axes[0]
)

# Different ellipse standard deviations
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    ellipse_std_devs=[2],
    title='2σ Ellipses Only',
    ax=axes[1]
)

plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    ellipse_std_devs=[1, 2, 3],
    title='1σ, 2σ and 3σ Ellipses',
    ax=axes[2]
)

plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    ellipse_std_devs=[0.5, 1, 1.5, 2, 2.5, 3],
    title='Multiple Ellipse Levels',
    ax=axes[3]
)

plt.tight_layout()
plt.show()

nrows, ncols = 2, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('GMM Visualization with Different Ellipse Options') # Basic GMM with ellipses plot_gmm( X, gmm=gmm, color_by_cluster=True, title='GMM with Ellipses (Default)', ax=axes[0] ) # Different ellipse standard deviations plot_gmm( X, gmm=gmm, color_by_cluster=True, ellipse_std_devs=[2], title='2σ Ellipses Only', ax=axes[1] ) plot_gmm( X, gmm=gmm, color_by_cluster=True, ellipse_std_devs=[1, 2, 3], title='1σ, 2σ and 3σ Ellipses', ax=axes[2] ) plot_gmm( X, gmm=gmm, color_by_cluster=True, ellipse_std_devs=[0.5, 1, 1.5, 2, 2.5, 3], title='Multiple Ellipse Levels', ax=axes[3] ) plt.tight_layout() plt.show()

Ellipse Styling Options¶

In [7]:

nrows, ncols = 3, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()

# Different ellipse styling options
ellipse_styles = [
    {'ellipse_fill': True, 'ellipse_alpha': 0.3, 'title': 'Filled Ellipses'},
    {'ellipse_fill': False, 'ellipse_line_width': 3, 'title': 'Outline Only'},
    {'ellipse_line_style': 'dashed', 'ellipse_line_width': 2, 'title': 'Dashed Lines'},
    {'ellipse_line_color': 'red', 'ellipse_line_alpha': 1.0, 'title': 'Red Borders'},
    {'ellipse_colors': 'Set3', 'title': 'Different Ellipse Colors'},
    {'ellipse_alpha': 0.8, 'ellipse_line_width': 4, 'ellipse_line_color': 'black', 'title': 'High Contrast'}
]

for i, style_options in enumerate(ellipse_styles):
    title = style_options.pop('title')
    plot_gmm(
        X,
        gmm=gmm,
        color_by_cluster=True,
        ellipse_std_devs=[1, 2],
        title=title,
        ax=axes[i],
        **style_options
    )

plt.tight_layout()
plt.show()

nrows, ncols = 3, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() # Different ellipse styling options ellipse_styles = [ {'ellipse_fill': True, 'ellipse_alpha': 0.3, 'title': 'Filled Ellipses'}, {'ellipse_fill': False, 'ellipse_line_width': 3, 'title': 'Outline Only'}, {'ellipse_line_style': 'dashed', 'ellipse_line_width': 2, 'title': 'Dashed Lines'}, {'ellipse_line_color': 'red', 'ellipse_line_alpha': 1.0, 'title': 'Red Borders'}, {'ellipse_colors': 'Set3', 'title': 'Different Ellipse Colors'}, {'ellipse_alpha': 0.8, 'ellipse_line_width': 4, 'ellipse_line_color': 'black', 'title': 'High Contrast'} ] for i, style_options in enumerate(ellipse_styles): title = style_options.pop('title') plot_gmm( X, gmm=gmm, color_by_cluster=True, ellipse_std_devs=[1, 2], title=title, ax=axes[i], **style_options ) plt.tight_layout() plt.show()

Log Probability Visualization¶

In [8]:

nrows, ncols = 2, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Continuous Coloring with Log Probabilities')

# Different continuous coloring options
colormaps = ['viridis', 'plasma', 'coolwarm', 'RdYlBu']

for i, cmap in enumerate(colormaps):
    plot_gmm(
        X,
        gmm=gmm,
        log_probs=log_probs,
        colormap=cmap,
        colorbar_label='Log Probability',
        title=f'Log Probabilities ({cmap})',
        show_ellipses=False,
        ax=axes[i]
    )

plt.tight_layout()
plt.show()

nrows, ncols = 2, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Continuous Coloring with Log Probabilities') # Different continuous coloring options colormaps = ['viridis', 'plasma', 'coolwarm', 'RdYlBu'] for i, cmap in enumerate(colormaps): plot_gmm( X, gmm=gmm, log_probs=log_probs, colormap=cmap, colorbar_label='Log Probability', title=f'Log Probabilities ({cmap})', show_ellipses=False, ax=axes[i] ) plt.tight_layout() plt.show()

Mean Markers and Initial Means¶

In [9]:

nrows, ncols = 2, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Component Means and Initial Means Visualization')

# Different mean marker options
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    show_means=True,
    mean_marker='X',
    mean_size=100,
    mean_color='black',
    title='Large X Markers for Means',
    ax=axes[0]
)

plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    show_means=True,
    mean_marker='*',
    mean_size=150,
    mean_color='white',
    title='White Star Markers',
    ax=axes[1]
)

# Show initial means if available
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    show_means=True,
    show_initial_means=True,
    initial_mean_marker='D',
    initial_mean_size=50,
    initial_mean_color='red',
    title='Final (black) and Initial (red) Means',
    ax=axes[2]
)

plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    show_means=False,
    show_ellipses=True,
    ellipse_alpha=0.2,
    title='Ellipses Only (No Mean Markers)',
    ax=axes[3]
)

plt.tight_layout()
plt.show()

nrows, ncols = 2, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Component Means and Initial Means Visualization') # Different mean marker options plot_gmm( X, gmm=gmm, color_by_cluster=True, show_means=True, mean_marker='X', mean_size=100, mean_color='black', title='Large X Markers for Means', ax=axes[0] ) plot_gmm( X, gmm=gmm, color_by_cluster=True, show_means=True, mean_marker='*', mean_size=150, mean_color='white', title='White Star Markers', ax=axes[1] ) # Show initial means if available plot_gmm( X, gmm=gmm, color_by_cluster=True, show_means=True, show_initial_means=True, initial_mean_marker='D', initial_mean_size=50, initial_mean_color='red', title='Final (black) and Initial (red) Means', ax=axes[2] ) plot_gmm( X, gmm=gmm, color_by_cluster=True, show_means=False, show_ellipses=True, ellipse_alpha=0.2, title='Ellipses Only (No Mean Markers)', ax=axes[3] ) plt.tight_layout() plt.show()

Component Weight Visualization¶

In [10]:

nrows, ncols = 3, 1
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Weight-Based Scaling of Ellipses')

# Regular visualization
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    title='No Weight Scaling',
    ax=axes[0]
)

# Scale alpha by weight
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    scale_alpha_by_weight=True,
    title='Alpha Scaled by Weight',
    ax=axes[1]
)

# Scale size by weight
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    scale_size_by_weight=True,
    title='Mean Size Scaled by Weight',
    ax=axes[2],
    mean_size=80,
)

plt.tight_layout()
plt.show()

# Print component weights for reference
print(f"Component weights: {gmm.weights_.detach().cpu().numpy()}")

nrows, ncols = 3, 1 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Weight-Based Scaling of Ellipses') # Regular visualization plot_gmm( X, gmm=gmm, color_by_cluster=True, title='No Weight Scaling', ax=axes[0] ) # Scale alpha by weight plot_gmm( X, gmm=gmm, color_by_cluster=True, scale_alpha_by_weight=True, title='Alpha Scaled by Weight', ax=axes[1] ) # Scale size by weight plot_gmm( X, gmm=gmm, color_by_cluster=True, scale_size_by_weight=True, title='Mean Size Scaled by Weight', ax=axes[2], mean_size=80, ) plt.tight_layout() plt.show() # Print component weights for reference print(f"Component weights: {gmm.weights_.detach().cpu().numpy()}")

Component weights: [0.06534623 0.330355   0.33683044 0.2674684 ]

Prediction Accuracy Assessment¶

In [11]:

nrows, ncols = 3, 1
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Prediction Accuracy Comparison')

# Show true labels
plot_gmm(
    X,
    gmm=None,
    color_by_cluster=True,
    true_labels=labels,
    title='True Labels',
    show_ellipses=False,
    ax=axes[0]
)

# Show predicted labels
plot_gmm(
    X,
    gmm=gmm,
    color_by_cluster=True,
    title='Predicted Labels',
    show_ellipses=False,
    ax=axes[1]
)

# Show correct vs incorrect predictions
plot_gmm(
    X,
    gmm=gmm,
    true_labels=labels,
    show_incorrect_predictions=True,
    match_labels_to_true=True,
    title='Correct (Green) vs Incorrect (Red)',
    show_ellipses=False,
    ax=axes[2]
)

plt.tight_layout()
plt.show()

nrows, ncols = 3, 1 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Prediction Accuracy Comparison') # Show true labels plot_gmm( X, gmm=None, color_by_cluster=True, true_labels=labels, title='True Labels', show_ellipses=False, ax=axes[0] ) # Show predicted labels plot_gmm( X, gmm=gmm, color_by_cluster=True, title='Predicted Labels', show_ellipses=False, ax=axes[1] ) # Show correct vs incorrect predictions plot_gmm( X, gmm=gmm, true_labels=labels, show_incorrect_predictions=True, match_labels_to_true=True, title='Correct (Green) vs Incorrect (Red)', show_ellipses=False, ax=axes[2] ) plt.tight_layout() plt.show()

Overview of Different Visualization Combinations¶

In [12]:

nrows, ncols = 2, 2
figsize = dynamic_figsize(nrows, ncols)
fig, axes = plt.subplots(nrows, ncols, figsize=figsize)
axes = axes.flatten()
fig.suptitle('Comprehensive GMM Visualization Examples')

# Custom legend labels
custom_labels = ['Cluster A', 'Cluster B', 'Cluster C', 'Cluster D']

# Full-featured visualization
plot_gmm(
    X,
    gmm=gmm,
    
    # Data points
    show_points=True,
    point_size=10,
    point_alpha=0.6,
    point_color='auto',
    
    # Clustering
    color_by_cluster=True,
    cluster_colors='Set2',
    legend_labels=custom_labels,
    
    # Ellipses
    show_ellipses=True,
    ellipse_std_devs=[1, 2, 3],
    ellipse_alpha=0.4,
    ellipse_fill=True,
    ellipse_line_style='solid',
    ellipse_line_width=1.5,
    ellipse_line_color='darkblue',
    ellipse_line_alpha=0.8,
    
    # Means
    show_means=True,
    mean_marker='*',
    mean_size=100,
    mean_color='black',
    
    # Initial means
    show_initial_means=True,
    initial_mean_marker='D',
    initial_mean_size=60,
    initial_mean_color='red',
    
    # Styling
    title='Complete GMM Visualization',
    xlabel='Feature Dimension 1',
    ylabel='Feature Dimension 2',
    ax=axes[0]
)

# Minimal visualization
plot_gmm(
    X,
    gmm=gmm,
    show_points=True,
    point_size=5,
    point_alpha=0.8,
    point_color='black',
    show_ellipses=False,
    show_means=False,
    legend=False,
    title='Minimal: Points Only',
    ax=axes[1]
)

# Ellipses only
plot_gmm(
    X,
    gmm=gmm,
    show_points=False,
    show_ellipses=True,
    ellipse_std_devs=[1, 2, 3],
    ellipse_alpha=0.6,
    ellipse_colors='viridis',
    show_means=True,
    mean_marker='X',
    mean_size=80,
    title='Ellipses and Means Only',
    ax=axes[2]
)

# Continuous coloring with ellipses
plot_gmm(
    X,
    gmm=gmm,
    log_probs=log_probs,
    colormap='coolwarm',
    colorbar_label='Log-Likelihood',
    show_ellipses=True,
    ellipse_std_devs=[2],
    ellipse_fill=False,
    ellipse_line_width=3,
    ellipse_line_color='black',
    title='Log Probabilities + Ellipses',
    ax=axes[3]
)

plt.tight_layout()
plt.show()

nrows, ncols = 2, 2 figsize = dynamic_figsize(nrows, ncols) fig, axes = plt.subplots(nrows, ncols, figsize=figsize) axes = axes.flatten() fig.suptitle('Comprehensive GMM Visualization Examples') # Custom legend labels custom_labels = ['Cluster A', 'Cluster B', 'Cluster C', 'Cluster D'] # Full-featured visualization plot_gmm( X, gmm=gmm, # Data points show_points=True, point_size=10, point_alpha=0.6, point_color='auto', # Clustering color_by_cluster=True, cluster_colors='Set2', legend_labels=custom_labels, # Ellipses show_ellipses=True, ellipse_std_devs=[1, 2, 3], ellipse_alpha=0.4, ellipse_fill=True, ellipse_line_style='solid', ellipse_line_width=1.5, ellipse_line_color='darkblue', ellipse_line_alpha=0.8, # Means show_means=True, mean_marker='*', mean_size=100, mean_color='black', # Initial means show_initial_means=True, initial_mean_marker='D', initial_mean_size=60, initial_mean_color='red', # Styling title='Complete GMM Visualization', xlabel='Feature Dimension 1', ylabel='Feature Dimension 2', ax=axes[0] ) # Minimal visualization plot_gmm( X, gmm=gmm, show_points=True, point_size=5, point_alpha=0.8, point_color='black', show_ellipses=False, show_means=False, legend=False, title='Minimal: Points Only', ax=axes[1] ) # Ellipses only plot_gmm( X, gmm=gmm, show_points=False, show_ellipses=True, ellipse_std_devs=[1, 2, 3], ellipse_alpha=0.6, ellipse_colors='viridis', show_means=True, mean_marker='X', mean_size=80, title='Ellipses and Means Only', ax=axes[2] ) # Continuous coloring with ellipses plot_gmm( X, gmm=gmm, log_probs=log_probs, colormap='coolwarm', colorbar_label='Log-Likelihood', show_ellipses=True, ellipse_std_devs=[2], ellipse_fill=False, ellipse_line_width=3, ellipse_line_color='black', title='Log Probabilities + Ellipses', ax=axes[3] ) plt.tight_layout() plt.show()