In [1]:

# Tratamiento de datos
# ==============================================================================
import pandas as pd
import numpy as np

# Preprocesado y modelado
# ==============================================================================
from sklearn.svm import SVC
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

# Configuración warnings
# ==============================================================================
import warnings
warnings.filterwarnings('ignore')

In [2]:

#Cargamos los datos!
url = 'https://raw.githubusercontent.com/JoaquinAmatRodrigo/' \
       + 'Estadistica-machine-learning-python/master/data/ESL.mixture.csv'
datos = pd.read_csv(url)
datos.head()

Out[2]:

	X1	X2
0	2.526093	0.321050
1	0.366954	0.031462
2	0.768219	0.717486
3	0.693436	0.777194
4	-0.019837	0.867254

In [3]:

#Visualizacion!
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(6,4))
ax.scatter(datos.X1, datos.X2, c=datos.y);
ax.set_title("Datos");

In [4]:

# División de los datos en train y test
X = datos.drop(columns = 'y')
y = datos['y']

In [5]:

Out[5]:

	X1	X2
0	2.526093	0.321050
1	0.366954	0.031462
2	0.768219	0.717486
3	0.693436	0.777194
4	-0.019837	0.867254
...	...	...
195	0.256750	2.293605
196	1.925173	0.165053
197	1.301941	0.992200
198	0.008131	2.242264
199	-0.196246	0.551404

200 rows × 2 columns

In [6]:

Out[6]:

0      0
1      0
2      0
3      0
4      0
      ..
195    1
196    1
197    1
198    1
199    1
Name: y, Length: 200, dtype: int64

In [7]:

X_train, X_test, y_train, y_test = train_test_split(X,y.values.reshape(-1,1),train_size= 0.7,random_state = 42,shuffle=True)

In [8]:

# Creación del modelo SVM 
modelo = SVC(C = 100, kernel = 'linear', random_state=42)
modelo.fit(X_train, y_train)

Out[8]:

SVC(C=100, kernel='linear', random_state=42)

In [9]:

#Predicciones!
y_test_pred = modelo.predict(X_test)

A lo largo de este notebook, se solicita calcular las métricas requeridas como así también su correspondiente interpretación:

Calcular la métrica Accuracy.

In [10]:

###Completar
from sklearn.metrics import accuracy_score
accuracy_score(y_test,y_test_pred)

Out[10]:

0.7

Crear la Matriz de Confusión

In [11]:

###Completar
from sklearn.metrics import confusion_matrix
confusion_matrix(y_test, y_test_pred) 

Out[11]:

array([[21, 10],
       [ 8, 21]])

Calcular la métrica F1 score

In [12]:

###Completar
from sklearn.metrics import f1_score
f1_score(y_test, y_test_pred) 

Out[12]:

0.7

Calcular todas las metricas al tiempo

In [13]:

from sklearn.metrics import classification_report
reporte=classification_report(y_test,y_test_pred)
print(reporte)

              precision    recall  f1-score   support

           0       0.72      0.68      0.70        31
           1       0.68      0.72      0.70        29

    accuracy                           0.70        60
   macro avg       0.70      0.70      0.70        60
weighted avg       0.70      0.70      0.70        60

Created in Deepnote