3016498
/
ANLP_WS24_CA2
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge branch 'main' of https://gitty.informatik.hs-mannheim.de/3016498/ANLP_WS24_CA2

main
arman 2025-02-16 00:43:10 +01:00
parent 95216088e5 64b2d5e411
commit 603eab83b4
3 changed files with 219 additions and 47 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
ml_plots.py
View File

@ -11,23 +11,23 @@ def save_plot(plt, plot_name):
time_stamp = time.strftime('%Y%m%d-%H%M%S')
plt.savefig(f'plots/{plot_name}_{time_stamp}.png')
def plot_training_history(hist_data, title='Training History', save=True):
def plot_training_history(hist_data, colors, title='Training History', save=True):
epochs = range(1, len(hist_data['train_loss']) + 1)
fig, axs = plt.subplots(1, 2, figsize=(12, 5))
# Plot accuracy
axs[1].plot(epochs, hist_data['train_rmse'], label='Train RMSE')
axs[1].plot(epochs, hist_data['val_rmse'], label='Validation RMSE')
axs[1].plot(epochs, hist_data['train_rmse'], label='Train RMSE', color=colors['blue'])
axs[1].plot(epochs, hist_data['val_rmse'], label='Validation RMSE', color=colors['green'])
axs[1].set_title('RMSE')
axs[1].set_xlabel('Epochs')
axs[1].set_ylabel('RMSE')
axs[1].legend()
# Plot loss
axs[0].plot(epochs, hist_data['train_loss'], label='Train Loss')
axs[0].plot(epochs, hist_data['val_loss'], label='Validation Loss')
axs[0].plot(epochs, hist_data['train_loss'], label='Train Loss', color=colors['blue'])
axs[0].plot(epochs, hist_data['val_loss'], label='Validation Loss', color=colors['green'])
axs[0].set_title('Loss')
axs[0].set_xlabel('Epochs')
axs[0].set_ylabel('Loss')
@ -41,10 +41,10 @@ def plot_training_history(hist_data, title='Training History', save=True):
save_plot(plt, title)
return plt
def plot_distribution(true_values, predicted_values, title='Distribution of Predicted and True Values', save=True):
def plot_distribution(true_values, predicted_values, colors, title='Distribution of Predicted and True Values', save=True):
plt.figure(figsize=(10, 6))
plt.hist(true_values, bins=20, color='skyblue', edgecolor='black', alpha=0.7, label='True Values')
plt.hist(predicted_values, bins=20, color='salmon', edgecolor='black', alpha=0.7, label='Predicted Values')
plt.hist(true_values, bins=20, color=colors['green'], edgecolor='black', alpha=0.7, label='True Values')
plt.hist(predicted_values, bins=20, color=colors['blue'], edgecolor='black', alpha=0.7, label='Predicted Values')
plt.title(title)
plt.xlabel('Score')
plt.ylabel('Frequency')
@ -55,15 +55,15 @@ def plot_distribution(true_values, predicted_values, title='Distribution of Pred
save_plot(plt, title)
return plt
def plot_predictions(true_values, predicted_values, title='True vs Predicted Values', threshold=0.3, save=True):
def plot_predictions(true_values, predicted_values, colors, title='True vs Predicted Values', threshold=0.3, save=True):
plt.figure(figsize=(10, 6))
# Difference between predicted and true values
correct_indices = np.isclose(true_values, predicted_values, atol=threshold)
incorrect_indices = ~correct_indices
# Plot
plt.scatter(np.array(true_values)[correct_indices], np.array(predicted_values)[correct_indices], color='green', label='Correctly predicted')
plt.scatter(np.array(true_values)[incorrect_indices], np.array(predicted_values)[incorrect_indices], color='red', label='Incorrectly predicted')
plt.plot([min(true_values), max(true_values)], [min(true_values), max(true_values)], color='blue', linestyle='--', label='Ideal Line')
plt.scatter(np.array(true_values)[correct_indices], np.array(predicted_values)[correct_indices], color=colors['green'], alpha=0.5, label='Correctly predicted')
plt.scatter(np.array(true_values)[incorrect_indices], np.array(predicted_values)[incorrect_indices], color=colors['red'], alpha=0.5, label='Incorrectly predicted')
plt.plot([min(true_values), max(true_values)], [min(true_values), max(true_values)], color=colors['blue'], linestyle='--', label='Ideal Line')
plt.xlabel('True Values')
plt.ylabel('Predicted Values')
plt.title(title)

14
model_comparison.ipynb
View File

@ -3,11 +3,7 @@
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"vscode": {
"languageId": "plaintext"
}
},
"metadata": {},
"outputs": [],
"source": [
"# TODO: compare"
@ -15,8 +11,14 @@
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python"
"name": "python",
"version": "3.10.4"
}
},
"nbformat": 4,

228
model_evaluation.ipynb
View File

File diff suppressed because one or more lines are too long