¶

  1. Implement Linear Regression, Ridge Regression and Lasso regression on teams dataset
  2. Use cross validation score and RMSE, R2 score.
  3. compare the results of various regression techniques
  4. Finally write your analysis

Linear Regression¶

In [2]:
import numpy as np 
import pandas as pd
In [3]:
data = pd.read_csv('teams.csv')
In [4]:
X = data.iloc[:,1:8]
Y = data['medals']
In [5]:
data.isnull().sum()
Out[5]:
team           0
year           0
athletes       0
events         0
age            0
height         0
weight         0
prev_medals    0
medals         0
dtype: int64
In [6]:
X
Out[6]:
year athletes events age height weight prev_medals
0 1964 8 8 22.0 161.0 64.2 0.0
1 1968 5 5 23.2 170.2 70.0 0.0
2 1972 8 8 29.0 168.3 63.8 0.0
3 1980 11 11 23.6 168.4 63.2 0.0
4 2004 5 5 18.6 170.8 64.8 0.0
... ... ... ... ... ... ... ...
2009 2000 26 19 25.0 179.0 71.1 0.0
2010 2004 14 11 25.1 177.8 70.5 0.0
2011 2008 16 15 26.1 171.9 63.7 3.0
2012 2012 9 8 27.3 174.4 65.2 4.0
2013 2016 31 13 27.5 167.8 62.2 0.0

2014 rows × 7 columns

In [7]:
Y
Out[7]:
0       0
1       0
2       0
3       0
4       0
       ..
2009    0
2010    3
2011    4
2012    0
2013    0
Name: medals, Length: 2014, dtype: int64
In [8]:
import pandas as pd
one_hot_encoded = pd.get_dummies(data, columns = ['team'])
In [9]:
from sklearn.model_selection import train_test_split

X_train, X_test, Y_train, Y_test =  train_test_split(X, Y, test_size = 0.3)

print(X_train.shape)
print(X_test.shape)
print(Y_train.shape)
print(Y_test.shape)

(1409, 7)
(605, 7)
(1409,)
(605,)
In [10]:
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error

lin_model = LinearRegression()
lin_model.fit(X_train, Y_train)
Out[10]:
LinearRegression()
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.
LinearRegression()
In [11]:
from sklearn.metrics import r2_score
In [12]:
y_test_predict = lin_model.predict(X_test)
In [13]:
rmse = (np.sqrt(mean_squared_error(Y_test, y_test_predict)))
r2 = r2_score(Y_test, y_test_predict)

print('RMSE is {}'.format(rmse))
print('R2 score is {}'.format(r2))

RMSE is 10.270951521656123
R2 score is 0.8543923284524518
In [14]:
from sklearn.linear_model import Lasso
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import RepeatedKFold
from sklearn.linear_model import Ridge
import pandas as pd
import numpy as np

data = pd.read_csv('teams.csv')
X = data.iloc[:,1:8]
Y = data['medals']

lr_model = LinearRegression()

cv = RepeatedKFold(n_splits=10, n_repeats=3, random_state=1)

scores = cross_val_score(lr_model, X, Y, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)

scores = np.absolute(scores)
print('Mean MAE: %.3f (%.3f)' % (np.mean(scores), np.std(scores)))

Mean MAE: 4.730 (0.597)

Lasso Regression¶

In [15]:
import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')

from sklearn.model_selection import cross_val_score
from sklearn.model_selection import RepeatedKFold
from sklearn.linear_model import Lasso

data = pd.read_csv('teams.csv')
X = data.iloc[:,1:8]
Y = data['medals']
In [16]:
lasso_model = Lasso(alpha=1.0)

cv = RepeatedKFold(n_splits=10, n_repeats=3, random_state=1)

scores = cross_val_score(lasso_model, X, Y, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)

scores = np.absolute(scores)
print('Mean MAE: %.3f (%.3f)' % (np.mean(scores), np.std(scores)))

Mean MAE: 4.657 (0.604)

Ridge Regression¶

In [17]:
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import RepeatedKFold
from sklearn.linear_model import Ridge
import pandas as pd
import numpy as np

data = pd.read_csv('teams.csv')
X = data.iloc[:,1:8]
Y = data['medals']

ridge_model = Ridge(alpha=1.0)

cv = RepeatedKFold(n_splits=10, n_repeats=3, random_state=1)

scores = cross_val_score(ridge_model, X, Y, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)

scores = np.absolute(scores)
print('Mean MAE: %.3f (%.3f)' % (np.mean(scores), np.std(scores)))

Mean MAE: 4.730 (0.597)
In [18]:
from sklearn.model_selection import KFold
from sklearn.model_selection import cross_val_score
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso
from sklearn.linear_model import Ridge
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split

data = pd.read_csv('teams.csv')
X = data.iloc[:,1:8]
Y = data['medals']

X_train, X_test, Y_train, Y_test =  train_test_split(X, Y, test_size = 0.3)

folds   = 10
metric  = "neg_mean_squared_error"


models = {}
models["Linear"]        = LinearRegression()
models["Lasso"]         = Lasso()
models['Ridge']         = Ridge()

model_results = []
model_names   = []
for model_name in models:
	model   = models[model_name]
	k_fold  = KFold(n_splits=folds)
	results = cross_val_score(model, X_train, Y_train, cv=k_fold, scoring=metric)
	
	model_results.append(results)
	model_names.append(model_name)
	print("{}: {}, {}".format(model_name, round(results.mean(), 2), round(results.std(), 2)))

Linear: -145.04, 76.22
Lasso: -144.98, 76.71
Ridge: -145.04, 76.22

Analysis:¶

Linear and Ridge regression are best suited for teams dataset based on mean values.