└── README.md


/README.md:
--------------------------------------------------------------------------------
 1 | # Multi-class-Classification-Logistic-Regression
 2 | 
 3 | The notebook is a complete **hands-on lab** for performing **multi-class classification** using a real-world dataset related to **obesity levels**. Here's a structured breakdown of its content:
 4 | 
 5 | ---
 6 | 
 7 | ### 🔍 **Main Purpose**
 8 | To demonstrate how to implement **multi-class classification strategies** in Python using scikit-learn on a labeled dataset about obesity.
 9 | 
10 | ---
11 | 
12 | ### 📂 **Dataset Used**
13 | - File: https://www.kaggle.com/datasets/ezzaldeenesmail/obesitydataset-raw-and-data-sinthetic
14 | - Loaded with: `pandas.read_csv()`
15 | - Target column: `NObeyesdad` (which represents obesity categories)
16 | 
17 | ---
18 | 
19 | ### 📌 **Notebook Structure**
20 | 
21 | #### 1. **Setup and Imports**
22 | ```python
23 | import numpy as np
24 | import pandas as pd
25 | import seaborn as sns
26 | import matplotlib.pylab as plt
27 | from sklearn.model_selection import train_test_split
28 | from sklearn.linear_model import LogisticRegression
29 | from sklearn.preprocessing import StandardScaler, OneHotEncoder
30 | from sklearn.multiclass import OneVsOneClassifier
31 | from sklearn.metrics import accuracy_score
32 | ```
33 | These are the main tools used for:
34 | - Data processing
35 | - Visualization
36 | - Training and evaluating classification models
37 | 
38 | ---
39 | 
40 | #### 2. **Data Loading & Exploration**
41 | - Load the dataset into a DataFrame
42 | - Display first few records (`data.head()`)
43 | - Visualize target variable distribution:
44 | ```python
45 | sns.countplot(y='NObeyesdad', data=data)
46 | ```
47 | 
48 | ---
49 | 
50 | #### 3. **Preprocessing**
51 | - Apply **One-Hot Encoding** for categorical variables
52 | - Use **StandardScaler** to normalize numerical features
53 | - Split data into **training** and **testing sets**
54 | 
55 | ---
56 | 
57 | #### 4. **Modeling**
58 | - Implements a **Logistic Regression** model using:
59 |   - **One-vs-Rest (OvR)** strategy
60 |   - **One-vs-One (OvO)** strategy
61 | 
62 | ```python
63 | # Example:
64 | model = LogisticRegression()
65 | ovo = OneVsOneClassifier(model)
66 | ovo.fit(X_train, y_train)
67 | ```
68 | 
69 | ---
70 | 
71 | #### 5. **Evaluation**
72 | - Evaluate model performance using **accuracy_score**
73 | - Compare results from OvO and OvR classifiers
74 | 
75 | ---
76 | 
77 | #### 6. **Conclusion & Summary**
78 | - Analyzes which strategy works better for the dataset
79 | - Shows final accuracy and insights
80 | 
81 | ---
82 | 
83 | Would you like a copy of the dataset used, a visual chart of the workflow, or a restructured version of the notebook for learning purposes?
84 | 


--------------------------------------------------------------------------------