In [1]:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

In [2]:

from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import confusion_matrix
from sklearn.metrics import ConfusionMatrixDisplay

In [3]:

from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import minmax_scale

In [4]:

messi_csv = pd.read_csv('messi_barca.csv', encoding = "ISO-8859-1", delimiter=',')  
messi_csv

Out[4]:

	Season	Age	Competition	Matches Played	Starts	Minutes played	Goals scored	Assists	PK	PKatt	CrdY	Goal/90	Ast/90	G+A/90	G-PK/90	G+A-PK/90
0	2004-2005	17	Champions League	1	1	90	0	0	0	0	0	0.00	0.00	0.00	0.00	0.00
1	2004-2005	17	La Liga	7	0	70	1	0	0	0	0	1.29	0.00	1.29	1.29	1.29
2	2005-2006	18	Champions League	6	4	322	1	1	0	0	0	0.28	0.28	0.56	0.28	0.56
3	2005-2006	18	La Liga	17	11	911	6	3	0	0	2	0.59	0.30	0.89	0.59	0.89
4	2006-2007	19	Champions League	5	4	385	1	0	0	0	1	0.23	0.00	0.23	0.23	0.23
5	2006-2007	19	La Liga	26	23	1983	14	3	0	0	2	0.64	0.14	0.77	0.64	0.77
6	2007-2008	20	Champions League	9	9	728	6	2	1	1	2	0.74	0.25	0.99	0.62	0.87
7	2007-2008	20	La Liga	27	23	1973	10	12	4	4	2	0.46	0.55	1.00	0.27	0.82
8	2008-2009	21	Champions League	12	10	927	9	5	0	0	1	0.87	0.49	1.36	0.87	1.36
9	2008-2009	21	La Liga	31	27	2516	23	11	3	4	2	0.82	0.39	1.22	0.72	1.11
10	2009-2010	22	Champions League	11	11	987	8	0	0	0	0	0.73	0.00	0.73	0.73	0.73
11	2009-2010	22	La Liga	35	30	2805	34	9	1	1	3	1.09	0.29	1.38	1.06	1.35
12	2010-2011	23	Champions League	13	11	1046	12	3	1	2	0	1.03	0.26	1.29	0.95	1.20
13	2010-2011	23	La Liga	33	31	2858	31	19	4	4	3	0.98	0.60	1.57	0.85	1.45
14	2011-2012	24	Champions League	11	11	990	14	5	4	5	2	1.27	0.45	1.73	0.91	1.36
15	2011-2012	24	La Liga	37	36	3270	50	16	10	11	6	1.38	0.44	1.82	1.10	1.54
16	2012-2013	25	Champions League	11	9	826	8	2	0	0	0	0.87	0.22	1.09	0.87	1.09
17	2012-2013	25	La Liga	32	28	2650	46	11	4	4	1	1.56	0.37	1.94	1.43	1.80
18	2013-2014	26	Champions League	7	7	630	8	0	2	2	0	1.14	0.00	1.14	0.86	0.86
19	2013-2014	26	La Liga	31	29	2501	28	11	7	8	2	1.01	0.40	1.40	0.76	1.15
20	2014-2015	27	Champions League	13	13	1147	10	6	0	1	1	0.78	0.47	1.26	0.78	1.26
21	2014-2015	27	Copa del Rey	6	6	540	5	3	1	2	1	0.83	0.50	1.33	0.67	1.17
22	2014-2015	27	La Liga	38	37	3375	43	18	5	6	4	1.15	0.48	1.63	1.01	1.49
23	2015-2016	28	Champions League	7	7	630	6	1	1	1	1	0.86	0.14	1.00	0.71	0.86
24	2015-2016	28	Copa del Rey	5	5	480	5	5	0	0	1	0.94	0.94	1.87	0.94	1.87
25	2015-2016	28	La Liga	33	31	2729	26	16	3	6	3	0.86	0.53	1.39	0.76	1.29
26	2015-2016	28	Supercopa de España	2	2	180	1	0	0	0	0	0.50	0.00	0.50	0.50	0.50
27	2016-2017	29	Champions League	9	9	810	11	2	2	2	0	1.22	0.22	1.44	1.00	1.22
28	2016-2017	29	Copa del Rey	7	7	630	5	3	1	1	3	0.71	0.43	1.14	0.57	1.00
29	2016-2017	29	La Liga	34	32	2828	37	9	6	7	6	1.18	0.29	1.46	0.99	1.27
30	2016-2017	29	Supercopa de España	2	2	180	1	2	0	0	0	0.50	1.00	1.50	0.50	1.50
31	2017-2018	30	Champions League	10	8	783	6	2	0	0	2	0.69	0.23	0.92	0.69	0.92
32	2017-2018	30	Copa del Rey	6	6	508	4	4	0	1	1	0.71	0.71	1.42	0.71	1.42
33	2017-2018	30	La Liga	36	32	3002	34	12	2	4	3	1.02	0.36	1.38	0.96	1.32
34	2017-2018	30	Supercopa de España	2	2	180	1	0	1	1	1	0.50	0.00	0.50	0.00	0.00
35	2018-2019	31	Champions League	10	9	838	12	3	1	1	0	1.29	0.32	1.61	1.18	1.50
36	2018-2019	31	Copa del Rey	5	4	388	3	2	0	0	0	0.70	0.46	1.16	0.70	1.16
37	2018-2019	31	La Liga	34	29	2713	36	13	4	5	3	1.19	0.43	1.63	1.06	1.49
38	2018-2019	31	Supercopa de España	1	1	90	0	1	0	0	0	0.00	1.00	1.00	0.00	1.00
39	2019-2020	32	Champions League	8	7	662	3	3	0	0	2	0.41	0.41	0.82	0.41	0.82
40	2019-2020	32	Copa del Rey	2	2	180	2	1	0	0	1	1.00	0.50	1.50	1.00	1.50
41	2019-2020	32	La Liga	33	32	2880	25	20	5	5	4	0.78	0.62	1.41	0.62	1.25
42	2019-2020	32	Supercopa de España	1	1	90	1	0	0	0	0	1.00	0.00	1.00	1.00	1.00

In [5]:

messi_csv.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 43 entries, 0 to 42
Data columns (total 17 columns):
 #   Column          Non-Null Count  Dtype  
---  ------          --------------  -----  
 0   Season          43 non-null     object 
 1   Age             43 non-null     int64  
 2   Competition     43 non-null     object 
 3   Matches Played  43 non-null     int64  
 4   Starts          43 non-null     int64  
 5   Minutes played  43 non-null     int64  
 6   Goals scored    43 non-null     int64  
 7   Assists         43 non-null     int64  
 8   PK              43 non-null     int64  
 9   PKatt           43 non-null     int64  
 10  CrdY            43 non-null     int64  
 11  CrdR            43 non-null     int64  
 12  Goal/90         43 non-null     float64
 13  Ast/90          43 non-null     float64
 14  G+A/90          43 non-null     float64
 15  G-PK/90         43 non-null     float64
 16  G+A-PK/90       43 non-null     float64
dtypes: float64(5), int64(10), object(2)
memory usage: 5.8+ KB

In [6]:

messi_csv.keys()

Out[6]:

Index(['Season', 'Age', 'Competition', 'Matches Played', 'Starts',
       'Minutes played', 'Goals scored', 'Assists', 'PK', 'PKatt', 'CrdY',
       'CrdR', 'Goal/90', 'Ast/90', 'G+A/90', 'G-PK/90', 'G+A-PK/90'],
      dtype='object')

KNN¶

In [7]:

X = np.array(messi_csv[['Age', 'Matches Played', 'Starts',
       'Minutes played', 'Goals scored', 'Assists', 'PK', 'PKatt', 'CrdY',
       'CrdR', 'Goal/90', 'Ast/90', 'G+A/90', 'G-PK/90', 'G+A-PK/90']])

In [8]:

competiciones = messi_csv['Competition']
comp = []
for c in competiciones:
    if c == 'Champions League':
        comp.append(0)
    if c == 'La Liga':
        comp.append(1)
    if c == 'Copa del Rey':
        comp.append(2)
    if c == 'Supercopa de España':
        comp.append(3)
        
Y = np.array(comp)

In [9]:

x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=1234, shuffle=True)

In [10]:

Out[10]:

array([0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 2,
       1, 0, 2, 1, 3, 0, 2, 1, 3, 0, 2, 1, 3, 0, 2, 1, 3, 0, 2, 1, 3])

In [11]:

knn = KNeighborsClassifier(5)

In [12]:

knn.fit(x_train, y_train)

Out[12]:

KNeighborsClassifier()

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

Matriz de confusion - validacion¶

In [13]:

y_pred = knn.predict(x_test)

In [14]:

y_pred

Out[14]:

array([0, 0, 1, 3, 0, 2, 1, 2, 0])

In [15]:

y_test

Out[15]:

array([0, 0, 1, 3, 0, 0, 1, 2, 0])

In [16]:

cm = confusion_matrix(y_test, y_pred)

In [17]:

cm_display = ConfusionMatrixDisplay(cm).plot()

Matriz de confusion - entrenamiento¶

In [18]:

y_pred = knn.predict(x_train)

In [19]:

y_pred

Out[19]:

array([0, 3, 1, 2, 0, 1, 1, 1, 3, 1, 1, 0, 0, 3, 0, 3, 3, 0, 1, 1, 1, 0,
       1, 0, 2, 0, 3, 3, 0, 1, 2, 0, 3, 1])

In [20]:

y_train

Out[20]:

array([0, 1, 1, 2, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 2, 3, 0, 1, 1, 1, 0,
       1, 2, 2, 0, 3, 3, 0, 1, 2, 0, 3, 1])

In [21]:

cm = confusion_matrix(y_train, y_pred)

In [22]:

cm_display = ConfusionMatrixDisplay(cm).plot()

Casos simplificado 2D¶

In [23]:

X = np.array(messi_csv[['Goals scored', 'Minutes played']])

In [24]:

x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=1234, shuffle=True)

In [25]:

knn.fit(x_train, y_train)

Out[25]:

KNeighborsClassifier()

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

In [26]:

y_pred = knn.predict(x_test)

In [27]:

knn.score(x_test, y_test)

Out[27]:

0.8888888888888888

In [28]:

knn.score(x_train, y_train)

Out[28]:

0.8235294117647058

In [29]:

cdict = {0: 'red', 1: 'green', 2: 'blue', 3: 'yellow'}
catdict= {0:'Champions League', 1: 'La Liga', 2:'Copa del Rey', 3:'Supercopa de España'}

figure, ax = plt.subplots(figsize=(8,8))

labels = Y
for g in np.unique(labels):
    ix = np.where(g == labels)
    ax.scatter(X[ix,0], X[ix,1], c = cdict[g], label = catdict[g], s = 45,edgecolor='k')

plt.xlabel('Goals Scored')
plt.ylabel('Minutes Played')
ax.legend()
plt.show()

In [30]:

# Plot the decision boundary. For that, we will assign a color to each
# point in the mesh [x_min, x_max]x[y_min, y_max].
x_min, x_max = X[:, 0].min() - 0.5,  X[:, 0].max() + 0.5
y_min, y_max =  X[:, 1].min() - 0.5,  X[:, 1].max() + 0.5
h = 1.0  # step size in the mesh
xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))
Z = knn.predict(np.c_[xx.ravel(), yy.ravel()])

# Put the result into a color plot
Z = Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx, yy, Z, cmap = "cool", shading ='auto')

# Plot also the training points
plt.scatter(x_train[:, 0], x_train[:, 1], c=y_train, edgecolors="k", cmap="jet",label='Training Points')
# Plot also the testing points
plt.scatter(x_test[:, 0], x_test[:, 1], c=y_pred, edgecolors="k", cmap="jet",marker="^",label='Test Points')
plt.legend(loc="upper left")
plt.xlabel("Matches played")
plt.ylabel("Goals Scored")

plt.show()

KNN escalado¶

In [31]:

X_scaled = minmax_scale(X)

In [32]:

x_train, x_test, y_train, y_test = train_test_split(X_scaled, Y, test_size=0.2, random_state=1234, shuffle=True)

In [33]:

knn = KNeighborsClassifier(5) # Jugar con este valor y ver la variacion en el resultado del grafico.

In [34]:

knn.fit(x_train, y_train)

Out[34]:

KNeighborsClassifier()

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

In [35]:

# Plot the decision boundary. For that, we will assign a color to each
# point in the mesh [x_min, x_max]x[y_min, y_max].
x_min, x_max = X_scaled[:, 0].min() - 0.5,  X_scaled[:, 0].max() + 0.5
y_min, y_max =  X_scaled[:, 1].min() - 0.5,  X_scaled[:, 1].max() + 0.5
h = 0.009  # step size in the mesh
xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))
Z = knn.predict(np.c_[xx.ravel(), yy.ravel()])

# Put the result into a color plot
Z = Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx, yy, Z, cmap = "cool", shading ='auto')

# Plot also the training points
plt.scatter(x_train[:, 0], x_train[:, 1], c=y_train, edgecolors="k", cmap="jet",label='Training Points')
# Plot also the testing points
plt.scatter(x_test[:, 0], x_test[:, 1], c=y_pred, edgecolors="k", cmap="jet",marker="^",label='Test Points')
plt.legend(loc="upper left")
plt.xlabel("Matches played")
plt.ylabel("Goals Scored")

plt.show()