Merge branch 'refactoring' of github.com:aimclub/FEDOT.LLM into refactoring

Author: AaLexUser

kaggle_comp/obesity_risks/output/report.md

@@ -51,7 +51,7 @@ Here’s a breakdown of the most critical sections of the code:
 
 1. **Data Preprocessing**  
    Converts the raw dataset to a clean format usable by the models.  
-   ```python
+```python
 from sklearn.impute import SimpleImputer
 from sklearn.preprocessing import MinMaxScaler
 
@@ -67,11 +67,11 @@ def transform_data(dataset: pd.DataFrame):
     features = scaler.fit_transform(features)
 
     return features, target
-   ```  
+```  
 
 2. **Model Training and AutoML**  
    Automated the model training process using the **FEDOT AutoML framework** to efficiently explore high-quality predictive models.  
-   ```python
+```python
 from fedot.api.main import Fedot
 from fedot.core.repository.tasks import Task, TaskTypesEnum
 
@@ -80,16 +80,16 @@ def train_model(features, target):
     pipeline = Fedot(problem='classification', preset='best_quality', timeout=1.0, metric='accuracy')
     pipeline.fit(features, target)
     return pipeline
-   ```  
+```  
 
 3. **Evaluation and Predictions**  
    Evaluated model performance and prepared submission files.  
-   ```python
+```python
 def create_submission():
     predictions = model.predict(test_features)
     submission_df = pd.DataFrame({'id': test_data['id'], 'NObeyesdad': predictions})
     submission_df.to_csv("submission.csv", index=False)
-   ```  
+```  
 
 ## 📊 Metrics  
 
@@ -105,4 +105,4 @@ This means the model correctly identifies the obesity risk category for 9 out of
 
 This model effectively predicts obesity risk with **90.7% accuracy**, a promising result for public health applications. By leveraging advanced AutoML techniques and robust preprocessing, it demonstrates a scalable, efficient approach to tackle similar classification problems. While this is a synthetic competition dataset, the pipeline could easily be adapted for real-world use cases like predicting cardiovascular risk or targeting dietary interventions.  
 
-Modeling obesity risk is not just about prediction—it's about enabling preventive healthcare measures that could save lives. This competition shows how machine learning can make serious strides in addressing global health challenges.  
+Modeling obesity risk is not just about prediction—it's about enabling preventive healthcare measures that could save lives. This competition shows how machine learning can make serious strides in addressing global health challenges.