diff --git a/.gitignore b/.gitignore
index c15e39e..7f6c528 100644
--- a/.gitignore
+++ b/.gitignore
@@ -6,4 +6,6 @@ env
*.pkl
*.ipynb
__pycache__
-.DS_Store
\ No newline at end of file
+.DS_Store
+tool_test_output.html
+tool_test_output.json
\ No newline at end of file
diff --git a/Dockerfile b/Dockerfile
index 68be808..987b676 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -9,7 +9,7 @@ RUN apt-get update && \
# Install Python packages
RUN pip install -U pip && \
- pip install --no-cache-dir --no-compile pycaret[models]==${VERSION} && \
+ pip install --no-cache-dir --no-compile pycaret[analysis,models]==${VERSION} && \
pip install --no-cache-dir --no-compile explainerdashboard
# Clean up unnecessary packages
diff --git a/tools/base_model_trainer.py b/tools/base_model_trainer.py
index 98f4009..d408905 100644
--- a/tools/base_model_trainer.py
+++ b/tools/base_model_trainer.py
@@ -2,19 +2,31 @@
import logging
import os
+from feature_importance import FeatureImportanceAnalyzer
+
import pandas as pd
+from utils import get_html_closing, get_html_template
+
logging.basicConfig(level=logging.DEBUG)
LOG = logging.getLogger(__name__)
class BaseModelTrainer:
- def __init__(self, input_file, target_col, output_dir, **kwargs):
+ def __init__(
+ self,
+ input_file,
+ target_col,
+ output_dir,
+ task_type,
+ **kwargs
+ ):
self.exp = None # This will be set in the subclass
self.input_file = input_file
self.target_col = target_col
self.output_dir = output_dir
+ self.task_type = task_type
self.data = None
self.target = None
self.best_model = None
@@ -29,9 +41,21 @@ def __init__(self, input_file, target_col, output_dir, **kwargs):
def load_data(self):
LOG.info(f"Loading data from {self.input_file}")
self.data = pd.read_csv(self.input_file, sep=None, engine='python')
+ self.data = self.data.apply(pd.to_numeric, errors='coerce')
names = self.data.columns.to_list()
self.target = names[int(self.target_col)-1]
- self.data = self.data.fillna(self.data.median(numeric_only=True))
+ if hasattr(self, 'missing_value_strategy'):
+ if self.missing_value_strategy == 'mean':
+ self.data = self.data.fillna(
+ self.data.mean(numeric_only=True))
+ elif self.missing_value_strategy == 'median':
+ self.data = self.data.fillna(
+ self.data.median(numeric_only=True))
+ elif self.missing_value_strategy == 'drop':
+ self.data = self.data.dropna()
+ else:
+ # Default strategy if not specified
+ self.data = self.data.fillna(self.data.median(numeric_only=True))
self.data.columns = self.data.columns.str.replace('.', '_')
def setup_pycaret(self):
@@ -116,113 +140,71 @@ def save_html_report(self):
setup_params_table = pd.DataFrame(
list(filtered_setup_params.items()),
columns=['Parameter', 'Value'])
- # Save model summary
+
best_model_params = pd.DataFrame(
self.best_model.get_params().items(),
columns=['Parameter', 'Value'])
best_model_params.to_csv(
os.path.join(self.output_dir, 'best_model.csv'),
index=False)
-
- # Save comparison results
self.results.to_csv(os.path.join(
self.output_dir, "comparison_results.csv"))
- # Read and encode plot images
plots_html = ""
for plot_name, plot_path in self.plots.items():
encoded_image = self.encode_image_to_base64(plot_path)
plots_html += f"""
-
PyCaret Model Training Report
+ {get_html_template()}
+
PyCaret Model Training Report
+
+
+ Setup & Best Model
+
+ Best Model Plots
+
+ Feature Importance
+
+
Setup Parameters
Parameter | Value |
- {setup_params_table.to_html(index=False,
- header=False, classes='table')}
+ {setup_params_table.to_html(
+ index=False, header=False, classes='table')}
Best Model: {model_name}
Parameter | Value |
- {best_model_params.to_html(index=False,
- header=False, classes='table')}
+ {best_model_params.to_html(
+ index=False, header=False, classes='table')}
Comparison Results
- {self.results.to_html(index=False,
- classes='table')}
+ {self.results.to_html(index=False, classes='table')}
-
Plots
+
+
+
Best Model Plots
{plots_html}
-
-
+
+ {feature_importance_html}
+
+ {get_html_closing()}
"""
with open(os.path.join(
diff --git a/tools/feature_importance.py b/tools/feature_importance.py
new file mode 100644
index 0000000..95e2e07
--- /dev/null
+++ b/tools/feature_importance.py
@@ -0,0 +1,175 @@
+import base64
+import logging
+import os
+
+import matplotlib.pyplot as plt
+
+import pandas as pd
+
+from pycaret.classification import ClassificationExperiment
+from pycaret.regression import RegressionExperiment
+
+logging.basicConfig(level=logging.DEBUG)
+LOG = logging.getLogger(__name__)
+
+
+class FeatureImportanceAnalyzer:
+ def __init__(
+ self,
+ task_type,
+ output_dir,
+ data_path=None,
+ data=None,
+ target_col=None):
+
+ if data is not None:
+ self.data = data
+ LOG.info("Data loaded from memory")
+ else:
+ self.target_col = target_col
+ self.data = pd.read_csv(data_path, sep=None, engine='python')
+ self.data.columns = self.data.columns.str.replace('.', '_')
+ self.data = self.data.fillna(self.data.median(numeric_only=True))
+ self.task_type = task_type
+ self.target = self.data.columns[int(target_col) - 1]
+ self.exp = ClassificationExperiment() \
+ if task_type == 'classification' \
+ else RegressionExperiment()
+ self.plots = {}
+ self.output_dir = output_dir
+
+ def setup_pycaret(self):
+ LOG.info("Initializing PyCaret")
+ setup_params = {
+ 'target': self.target,
+ 'session_id': 123,
+ 'html': True,
+ 'log_experiment': False,
+ 'system_log': False
+ }
+ LOG.info(self.task_type)
+ LOG.info(self.exp)
+ self.exp.setup(self.data, **setup_params)
+
+ def save_coefficients(self):
+ model = self.exp.create_model('lr')
+ coef_df = pd.DataFrame({
+ 'Feature': self.data.columns.drop(self.target),
+ 'Coefficient': model.coef_[0]
+ })
+ coef_html = coef_df.to_html(index=False)
+ return coef_html
+
+ def save_tree_importance(self):
+ model = self.exp.create_model('rf')
+ importances = model.feature_importances_
+ feature_importances = pd.DataFrame({
+ 'Feature': self.data.columns.drop(self.target),
+ 'Importance': importances
+ }).sort_values(by='Importance', ascending=False)
+ plt.figure(figsize=(10, 6))
+ plt.barh(
+ feature_importances['Feature'],
+ feature_importances['Importance'])
+ plt.xlabel('Importance')
+ plt.title('Feature Importance (Random Forest)')
+ plot_path = os.path.join(
+ self.output_dir,
+ 'tree_importance.png')
+ plt.savefig(plot_path)
+ plt.close()
+ self.plots['tree_importance'] = plot_path
+
+ def save_shap_values(self):
+ model = self.exp.create_model('lightgbm')
+ import shap
+ explainer = shap.Explainer(model)
+ shap_values = explainer.shap_values(
+ self.data.drop(columns=[self.target]))
+ shap.summary_plot(shap_values, self.data.drop(
+ columns=[self.target]), show=False)
+ plt.title('Shap (LightGBM)')
+ plot_path = os.path.join(
+ self.output_dir, 'shap_summary.png')
+ plt.savefig(plot_path)
+ plt.close()
+ self.plots['shap_summary'] = plot_path
+
+ def generate_feature_importance(self):
+ coef_html = self.save_coefficients()
+ self.save_tree_importance()
+ self.save_shap_values()
+ return coef_html
+
+ def encode_image_to_base64(self, img_path):
+ with open(img_path, 'rb') as img_file:
+ return base64.b64encode(img_file.read()).decode('utf-8')
+
+ def generate_html_report(self, coef_html):
+ LOG.info("Generating HTML report")
+
+ # Read and encode plot images
+ plots_html = ""
+ for plot_name, plot_path in self.plots.items():
+ encoded_image = self.encode_image_to_base64(plot_path)
+ plots_html += f"""
+
+
Feature importance analysis from a
+ trained Random Forest
+
{'Use gini impurity for'
+ 'calculating feature importance for classification'
+ 'and Variance Reduction for regression'
+ if plot_name == 'tree_importance'
+ else 'SHAP Summary from a trained lightgbm'}
+
+
+ """
+
+ # Generate HTML content with tabs
+ html_content = f"""
+
PyCaret Feature Importance Report
+
+
+
Coefficients (based on a trained
+ {'Logistic Regression'
+ if self.task_type == 'classification'
+ else 'Linear Regression'} Model)
+
{coef_html}
+
+ {plots_html}
+ """
+
+ return html_content
+
+ def run(self):
+ LOG.info("Running feature importance analysis")
+ self.setup_pycaret()
+ coef_html = self.generate_feature_importance()
+ html_content = self.generate_html_report(coef_html)
+ LOG.info("Feature importance analysis completed")
+ return html_content
+
+
+if __name__ == "__main__":
+ import argparse
+ parser = argparse.ArgumentParser(description="Feature Importance Analysis")
+ parser.add_argument(
+ "--data_path", type=str, help="Path to the dataset")
+ parser.add_argument(
+ "--target_col", type=int,
+ help="Index of the target column (1-based)")
+ parser.add_argument(
+ "--task_type", type=str,
+ choices=["classification", "regression"],
+ help="Task type: classification or regression")
+ parser.add_argument(
+ "--output_dir",
+ type=str,
+ help="Directory to save the outputs")
+ args = parser.parse_args()
+
+ analyzer = FeatureImportanceAnalyzer(
+ args.data_path, args.target_col,
+ args.task_type, args.output_dir)
+ analyzer.run()
diff --git a/tools/pycaret_classification.py b/tools/pycaret_classification.py
index 0ef935e..d994015 100644
--- a/tools/pycaret_classification.py
+++ b/tools/pycaret_classification.py
@@ -10,8 +10,15 @@
class ClassificationModelTrainer(BaseModelTrainer):
- def __init__(self, input_file, target_col, output_dir, **kwargs):
- super().__init__(input_file, target_col, output_dir, **kwargs)
+ def __init__(
+ self,
+ input_file,
+ target_col,
+ output_dir,
+ task_type,
+ **kwargs):
+ super().__init__(
+ input_file, target_col, output_dir, task_type, **kwargs)
self.exp = ClassificationExperiment()
def save_dashboard(self):
diff --git a/tools/pycaret_regression.py b/tools/pycaret_regression.py
index 91d5b7a..2f1f80e 100644
--- a/tools/pycaret_regression.py
+++ b/tools/pycaret_regression.py
@@ -10,8 +10,15 @@
class RegressionModelTrainer(BaseModelTrainer):
- def __init__(self, input_file, target_col, output_dir, **kwargs):
- super().__init__(input_file, target_col, output_dir, **kwargs)
+ def __init__(
+ self,
+ input_file,
+ target_col,
+ output_dir,
+ task_type,
+ **kwargs):
+ super().__init__(
+ input_file, target_col, output_dir, task_type, **kwargs)
self.exp = RegressionExperiment()
def save_dashboard(self):
diff --git a/tools/pycaret_train.py b/tools/pycaret_train.py
index 1f15c8f..534a997 100644
--- a/tools/pycaret_train.py
+++ b/tools/pycaret_train.py
@@ -85,13 +85,17 @@ def main():
trainer = ClassificationModelTrainer(
args.input_file,
args.target_col,
- args.output_dir, **model_kwargs)
+ args.output_dir,
+ args.model_type,
+ **model_kwargs)
elif args.model_type == "regression":
if "fix_imbalance" in model_kwargs:
del model_kwargs["fix_imbalance"]
trainer = RegressionModelTrainer(
args.input_file,
- args.target_col, args.output_dir,
+ args.target_col,
+ args.output_dir,
+ args.model_type,
**model_kwargs)
else:
LOG.error("Invalid model type. Please choose \
diff --git a/tools/pycaret_train.xml b/tools/pycaret_train.xml
index 2a06c11..06ed7bb 100644
--- a/tools/pycaret_train.xml
+++ b/tools/pycaret_train.xml
@@ -1,12 +1,12 @@
- Compare different machine learning models on a dataset using PyCaret.
+ Compare different machine learning models on a dataset using PyCaret. Do feature analysis using LR, Random Forest and LightGBM.
pycaret_macros.xml
@@ -152,10 +155,12 @@
-
+
-
+
-
+
This tool uses PyCaret to train and evaluate machine learning models.
- Ensure that the Conda environment specified in the requirements is correctly set up.
\ No newline at end of file
diff --git a/tools/test-data/auto-mpg.csv b/tools/test-data/auto-mpg.tsv
similarity index 100%
rename from tools/test-data/auto-mpg.csv
rename to tools/test-data/auto-mpg.tsv
diff --git a/tools/test-data/expected_best_model_classification.csv b/tools/test-data/expected_best_model_classification.csv
new file mode 100644
index 0000000..81152e5
--- /dev/null
+++ b/tools/test-data/expected_best_model_classification.csv
@@ -0,0 +1,20 @@
+Parameter,Value
+boosting_type,gbdt
+class_weight,
+colsample_bytree,1.0
+importance_type,split
+learning_rate,0.1
+max_depth,-1
+min_child_samples,20
+min_child_weight,0.001
+min_split_gain,0.0
+n_estimators,100
+n_jobs,-1
+num_leaves,31
+objective,
+random_state,123
+reg_alpha,0.0
+reg_lambda,0.0
+subsample,1.0
+subsample_for_bin,200000
+subsample_freq,0
diff --git a/tools/test-data/expected_best_model_regression.csv b/tools/test-data/expected_best_model_regression.csv
new file mode 100644
index 0000000..81152e5
--- /dev/null
+++ b/tools/test-data/expected_best_model_regression.csv
@@ -0,0 +1,20 @@
+Parameter,Value
+boosting_type,gbdt
+class_weight,
+colsample_bytree,1.0
+importance_type,split
+learning_rate,0.1
+max_depth,-1
+min_child_samples,20
+min_child_weight,0.001
+min_split_gain,0.0
+n_estimators,100
+n_jobs,-1
+num_leaves,31
+objective,
+random_state,123
+reg_alpha,0.0
+reg_lambda,0.0
+subsample,1.0
+subsample_for_bin,200000
+subsample_freq,0
diff --git a/tools/test-data/expected_comparison_result_classification.html b/tools/test-data/expected_comparison_result_classification.html
index 9f17520..da98cd0 100644
--- a/tools/test-data/expected_comparison_result_classification.html
+++ b/tools/test-data/expected_comparison_result_classification.html
@@ -1,62 +1,98 @@
-
-
-
-
-
-
PyCaret Model Training Report
-
-
-
-
-
PyCaret Model Training Report
+
+
+
+
Model Training Report
+
+
+
+
+
+
PyCaret Model Training Report
+
+
+ Setup & Best Model
+
+ Best Model Plots
+
+ Feature Importance
+
+
Setup Parameters
Parameter | Value |
@@ -183,7 +219,29 @@ Comparison Results
0.7074 |
0.4668 |
0.4969 |
- 0.042 |
+ 0.531 |
+
+
+ Extreme Gradient Boosting |
+ 0.735 |
+ 0.7000 |
+ 0.7500 |
+ 0.6433 |
+ 0.6717 |
+ 0.4289 |
+ 0.4523 |
+ 0.201 |
+
+
+ Logistic Regression |
+ 0.730 |
+ 0.7667 |
+ 0.6667 |
+ 0.5933 |
+ 0.6150 |
+ 0.4013 |
+ 0.4167 |
+ 0.080 |
Quadratic Discriminant Analysis |
@@ -194,18 +252,18 @@ Comparison Results
0.5933 |
0.4514 |
0.4929 |
- 0.027 |
+ 0.063 |
- Logistic Regression |
+ CatBoost Classifier |
0.730 |
- 0.7667 |
- 0.6667 |
- 0.5933 |
- 0.6150 |
- 0.4013 |
- 0.4167 |
- 0.025 |
+ 0.7333 |
+ 0.7500 |
+ 0.6600 |
+ 0.6783 |
+ 0.4293 |
+ 0.4521 |
+ 7.149 |
Gradient Boosting Classifier |
@@ -216,7 +274,7 @@ Comparison Results
0.6900 |
0.4117 |
0.4546 |
- 0.157 |
+ 0.295 |
Random Forest Classifier |
@@ -227,7 +285,7 @@ Comparison Results
0.6383 |
0.3783 |
0.4058 |
- 0.159 |
+ 0.392 |
Linear Discriminant Analysis |
@@ -238,7 +296,7 @@ Comparison Results
0.6717 |
0.3690 |
0.4135 |
- 0.029 |
+ 0.071 |
K Neighbors Classifier |
@@ -249,7 +307,7 @@ Comparison Results
0.6655 |
0.3634 |
0.3779 |
- 0.027 |
+ 0.101 |
Decision Tree Classifier |
@@ -260,7 +318,7 @@ Comparison Results
0.6574 |
0.3523 |
0.3854 |
- 0.025 |
+ 0.083 |
Naive Bayes |
@@ -271,7 +329,7 @@ Comparison Results
0.5917 |
0.3244 |
0.3333 |
- 0.029 |
+ 0.117 |
Ridge Classifier |
@@ -282,7 +340,7 @@ Comparison Results
0.6017 |
0.3320 |
0.3500 |
- 0.032 |
+ 0.062 |
Extra Trees Classifier |
@@ -293,7 +351,7 @@ Comparison Results
0.5805 |
0.2650 |
0.2816 |
- 0.111 |
+ 0.323 |
Ada Boost Classifier |
@@ -304,7 +362,7 @@ Comparison Results
0.5933 |
0.2697 |
0.3121 |
- 0.163 |
+ 0.276 |
SVM - Linear Kernel |
@@ -315,7 +373,7 @@ Comparison Results
0.4717 |
0.1306 |
0.1647 |
- 0.029 |
+ 0.063 |
Dummy Classifier |
@@ -326,84 +384,199 @@ Comparison Results
0.0000 |
0.0000 |
0.0000 |
- 0.033 |
+ 0.074 |
-
Plots
+
+
+
Best Model Plots
Auc
-
+
Confusion_matrix
-
+
Threshold
-
+
Pr
-
+
Error
-
+
Class_report
-
+
Learning
-
+
Calibration
-
+
Vc
-
+
Dimension
-
+
Manifold
-
+
Rfe
-
+
Feature
-
+
Feature_all
-
+
+
+
+
+
+
+
PyCaret Feature Importance Report
+
+
+
Coefficients (based on a trained
+ Logistic Regression Model)
+
+
+
+ Feature |
+ Coefficient |
+
+
+
+
+ SCGB2A2 |
+ -0.588679 |
+
+
+ FDCSP |
+ 0.772756 |
+
+
+ MUCL1 |
+ 0.168434 |
+
+
+ PIP |
+ 0.557828 |
+
+
+ TFF1 |
+ -0.989886 |
+
+
+ SCGB1D1 |
+ -0.740511 |
+
+
+ SCGB1D2 |
+ -0.727969 |
+
+
+ CALML5 |
+ 0.392070 |
+
+
+ AGR2 |
+ -0.716675 |
+
+
+ CPB1 |
+ 0.059255 |
+
+
+
+
+
+
+
Feature importance analysis from a
+ trained Random Forest
+
Use gini impurity forcalculating feature importance for classificationand Variance Reduction for regression
+
+
+
+
+
Feature importance analysis from a
+ trained Random Forest
+
SHAP Summary from a trained lightgbm
+
+
-
-
+
+
+
+
+
+
\ No newline at end of file
diff --git a/tools/test-data/expected_comparison_result_regression.html b/tools/test-data/expected_comparison_result_regression.html
index 7e78624..a3c38bd 100644
--- a/tools/test-data/expected_comparison_result_regression.html
+++ b/tools/test-data/expected_comparison_result_regression.html
@@ -1,63 +1,98 @@
-
-
-
-
-
-
PyCaret Model Training Report
-
-
-
-
-
PyCaret Model Training Report
+
+
+
+
Model Training Report
+
+
+
+
+
+
PyCaret Model Training Report
+
+
+ Setup & Best Model
+
+ Best Model Plots
+
+ Feature Importance
+
+
Setup Parameters
Parameter | Value |
@@ -74,31 +109,87 @@ Setup Parameters
-
Best Model: CatBoostRegressor
+
Best Model: LGBMRegressor
Parameter | Value |
- loss_function |
- RMSE |
+ boosting_type |
+ gbdt |
+
+
+ class_weight |
+ None |
+
+
+ colsample_bytree |
+ 1.0 |
+
+
+ importance_type |
+ split |
+
+
+ learning_rate |
+ 0.1 |
+
+
+ max_depth |
+ -1 |
+
+
+ min_child_samples |
+ 20 |
+
+
+ min_child_weight |
+ 0.001 |
+
+
+ min_split_gain |
+ 0.0 |
+
+
+ n_estimators |
+ 100 |
- border_count |
- 254 |
+ n_jobs |
+ -1 |
- verbose |
- False |
+ num_leaves |
+ 31 |
- task_type |
- CPU |
+ objective |
+ None |
random_state |
123 |
+
+ reg_alpha |
+ 0.0 |
+
+
+ reg_lambda |
+ 0.0 |
+
+
+ subsample |
+ 1.0 |
+
+
+ subsample_for_bin |
+ 200000 |
+
+
+ subsample_freq |
+ 0 |
+
@@ -118,175 +209,175 @@
Comparison Results
+
+ Light Gradient Boosting Machine |
+ 1.9064 |
+ 6.9812 |
+ 2.5932 |
+ 0.8852 |
+ 0.1012 |
+ 0.0832 |
+ 0.041 |
+
CatBoost Regressor |
- 2.0857 |
- 8.2406 |
- 2.8434 |
- 0.8616 |
- 0.1146 |
- 0.0932 |
- 8.470 |
+ 1.8984 |
+ 7.1071 |
+ 2.6242 |
+ 0.8818 |
+ 0.1017 |
+ 0.0827 |
+ 5.248 |
Extra Trees Regressor |
- 2.1028 |
- 8.2773 |
- 2.8615 |
- 0.8590 |
- 0.1122 |
- 0.0922 |
- 3.755 |
+ 1.8770 |
+ 7.1233 |
+ 2.6496 |
+ 0.8796 |
+ 0.1001 |
+ 0.0805 |
+ 0.169 |
- Light Gradient Boosting Machine |
- 2.1336 |
- 8.5588 |
- 2.9093 |
- 0.8578 |
- 0.1138 |
- 0.0930 |
- 1.483 |
+ Random Forest Regressor |
+ 1.9666 |
+ 7.6187 |
+ 2.7155 |
+ 0.8738 |
+ 0.1051 |
+ 0.0861 |
+ 0.284 |
- Random Forest Regressor |
- 2.2340 |
- 9.3259 |
- 3.0304 |
- 0.8438 |
- 0.1189 |
- 0.0984 |
- 5.435 |
+ Extreme Gradient Boosting |
+ 2.0811 |
+ 8.7754 |
+ 2.8962 |
+ 0.8576 |
+ 0.1087 |
+ 0.0896 |
+ 0.094 |
Gradient Boosting Regressor |
- 2.1911 |
- 9.4807 |
- 3.0328 |
- 0.8430 |
- 0.1178 |
- 0.0955 |
- 0.874 |
+ 2.0406 |
+ 8.9163 |
+ 2.9113 |
+ 0.8538 |
+ 0.1104 |
+ 0.0878 |
+ 0.130 |
- Extreme Gradient Boosting |
- 2.3047 |
- 10.3017 |
- 3.1834 |
- 0.8299 |
- 0.1219 |
- 0.1009 |
- 0.822 |
+ AdaBoost Regressor |
+ 2.2673 |
+ 10.1260 |
+ 3.1351 |
+ 0.8339 |
+ 0.1207 |
+ 0.1000 |
+ 0.119 |
- Elastic Net |
- 2.5336 |
- 11.3393 |
- 3.3316 |
- 0.8147 |
- 0.1398 |
- 0.1145 |
- 0.238 |
+ Ridge Regression |
+ 2.5315 |
+ 11.4346 |
+ 3.3421 |
+ 0.8120 |
+ 0.1461 |
+ 0.1173 |
+ 0.024 |
- Bayesian Ridge |
- 2.5427 |
- 11.5743 |
- 3.3639 |
+ Linear Regression |
+ 2.5325 |
+ 11.4367 |
+ 3.3424 |
0.8119 |
- 0.1401 |
- 0.1154 |
- 3.023 |
+ 0.1460 |
+ 0.1173 |
+ 0.029 |
- Lasso Least Angle Regression |
- 2.5705 |
- 11.6280 |
- 3.3736 |
- 0.8099 |
- 0.1422 |
- 0.1163 |
- 0.551 |
-
-
- Lasso Regression |
- 2.5706 |
- 11.6280 |
- 3.3736 |
- 0.8099 |
- 0.1422 |
- 0.1163 |
- 0.204 |
+ Bayesian Ridge |
+ 2.5238 |
+ 11.4695 |
+ 3.3477 |
+ 0.8113 |
+ 0.1480 |
+ 0.1173 |
+ 0.023 |
- Ridge Regression |
- 2.5765 |
- 11.7974 |
- 3.3966 |
- 0.8082 |
- 0.1418 |
- 0.1172 |
- 0.258 |
+ Least Angle Regression |
+ 2.6531 |
+ 12.2959 |
+ 3.4615 |
+ 0.7986 |
+ 0.1506 |
+ 0.1224 |
+ 0.025 |
- Linear Regression |
- 2.5809 |
- 11.8270 |
- 3.4009 |
- 0.8077 |
- 0.1420 |
- 0.1174 |
- 0.270 |
+ Elastic Net |
+ 2.6266 |
+ 12.3303 |
+ 3.4751 |
+ 0.7971 |
+ 0.1543 |
+ 0.1217 |
+ 0.024 |
- Least Angle Regression |
- 2.6615 |
- 12.3747 |
- 3.4723 |
- 0.7994 |
- 0.1458 |
- 0.1210 |
- 0.247 |
+ Huber Regressor |
+ 2.4866 |
+ 12.4597 |
+ 3.4839 |
+ 0.7967 |
+ 0.1387 |
+ 0.1104 |
+ 0.053 |
- AdaBoost Regressor |
+ Lasso Least Angle Regression |
2.6444 |
- 12.6052 |
- 3.5382 |
- 0.7885 |
- 0.1393 |
- 0.1175 |
- 1.196 |
+ 12.4441 |
+ 3.4920 |
+ 0.7954 |
+ 0.1547 |
+ 0.1224 |
+ 0.023 |
- Huber Regressor |
- 2.5126 |
- 13.0411 |
- 3.5739 |
- 0.7862 |
- 0.1355 |
- 0.1084 |
- 3.817 |
+ Lasso Regression |
+ 2.6446 |
+ 12.4444 |
+ 3.4921 |
+ 0.7954 |
+ 0.1547 |
+ 0.1224 |
+ 0.026 |
Decision Tree Regressor |
- 2.8325 |
- 15.5690 |
- 3.8947 |
- 0.7369 |
- 0.1522 |
- 0.1233 |
- 0.845 |
+ 2.7032 |
+ 13.3008 |
+ 3.6071 |
+ 0.7724 |
+ 0.1419 |
+ 0.1185 |
+ 0.030 |
K Neighbors Regressor |
- 3.2820 |
- 19.4715 |
- 4.3865 |
- 0.6744 |
- 0.1637 |
- 0.1389 |
- 0.783 |
+ 3.1884 |
+ 18.3559 |
+ 4.2627 |
+ 0.6902 |
+ 0.1604 |
+ 0.1351 |
+ 0.030 |
Orthogonal Matching Pursuit |
@@ -296,7 +387,7 @@ Comparison Results
0.6686 |
0.1755 |
0.1479 |
- 0.997 |
+ 0.021 |
Dummy Regressor |
@@ -306,78 +397,167 @@ Comparison Results
-0.0687 |
0.3355 |
0.3285 |
- 0.273 |
+ 0.030 |
Passive Aggressive Regressor |
- 11.7360 |
- 200.7733 |
- 13.3007 |
- -2.6451 |
- 0.6835 |
- 0.5873 |
- 0.644 |
+ 10.6628 |
+ 178.4730 |
+ 12.2226 |
+ -2.3674 |
+ 0.6393 |
+ 0.5210 |
+ 0.024 |
-
Plots
+
+
+
Best Model Plots
Residuals
-
+
Error
-
+
Cooks
-
+
Learning
-
+
Vc
-
+
Manifold
-
+
Rfe
-
+
Feature
-
+
Feature_all
+
+
+
+
+
+
+
PyCaret Feature Importance Report
+
+
+
Coefficients (based on a trained
+ Linear Regression Model)
+
+
+
+ Feature |
+ Coefficient |
+
+
+
+
+ Cylinders |
+ -0.414454 |
+
+
+ Displacement |
+ -0.414454 |
+
+
+ Horsepower |
+ -0.414454 |
+
+
+ Weight |
+ -0.414454 |
+
+
+ Acceleration |
+ -0.414454 |
+
+
+ ModelYear |
+ -0.414454 |
+
+
+ Origin |
+ -0.414454 |
+
+
+
+
+
+
+
Feature importance analysis from a
+ trained Random Forest
+
Use gini impurity forcalculating feature importance for classificationand Variance Reduction for regression
+
+
+
+
+
Feature importance analysis from a
+ trained Random Forest
+
SHAP Summary from a trained lightgbm
+ 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" alt="shap_summary">
+
-
-
+
+
+
+
+
+
\ No newline at end of file
diff --git a/tools/test-data/expected_dashboard_classification.html b/tools/test-data/expected_dashboard_classification.html
index 8e9497b..c89879c 100644
--- a/tools/test-data/expected_dashboard_classification.html
+++ b/tools/test-data/expected_dashboard_classification.html
@@ -58,7 +58,7 @@
@@ -179,7 +179,7 @@
@@ -213,7 +213,7 @@