week 5. Regression Analysis _ 현주

 

ch06.ipynb

0.01MB

 

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

In [ ]:

from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive

In [ ]:

ice = pd.read_csv('/content/drive/MyDrive/Colab Notebooks/SeaIce.txt',delim_whitespace=True)
print('shape:', ice.shape)

shape: (424, 6)

In [ ]:

ice2 = ice[ice.data_type != '-9999']
ice2

Out[ ]:

yearmodata_typeregionextentarea01234...419420421422423

1979	1	Goddard	N	15.54	12.33
1980	1	Goddard	N	14.96	11.85
1981	1	Goddard	N	15.03	11.82
1982	1	Goddard	N	15.26	12.11
1983	1	Goddard	N	15.10	11.92
...	...	...	...	...	...
2009	12	Goddard	N	12.51	10.25
2010	12	Goddard	N	12.02	10.08
2011	12	Goddard	N	12.40	10.28
2012	12	Goddard	N	12.20	10.11
2013	12	NRTSI-G	N	12.38	10.48

422 rows × 6 columns

 

In [ ]:

import seaborn as sns
sns.lmplot(x="mo", y="extent", data=ice2)

Out[ ]:

<seaborn.axisgrid.FacetGrid at 0x7db49c125150>

In [ ]:

grouped = ice2.groupby('mo')
month_means = grouped.extent.mean()
month_means

Out[ ]:

mo
1     14.479429
2     15.298889
3     15.491714
4     14.766000
5     13.396000
6     11.860000
7      9.601143
8      7.122286
9      6.404857
10     8.809143
11    10.964722
12    13.059429
Name: extent, dtype: float64

 

for i in range(12):
    ice2.extent[ice2.mo == i+1] = 100 * (ice2.extent[ice2.mo == i+1] - month_means[i+1]) /month_means.mean()
sns.lmplot(x="mo", y="extent", data=ice2)

 

Out[ ]:

<seaborn.axisgrid.FacetGrid at 0x7db46105a2f0>

sns.lmplot(x = "year", y = "extent", data = ice2)

Out[ ]:

<seaborn.axisgrid.FacetGrid at 0x7db49c125420>

In [ ]:

from sklearn.linear_model import LinearRegression
est = LinearRegression(fit_intercept = True)

In [ ]:

x = ice2[['year']]
y = ice2[['extent']]
est.fit(x, y)
print("Coefficients: ", est.coef_)
print("Intercept: ", est.intercept_)

Coefficients:  [[-0.45275459]]
Intercept:  [903.71640207]

In [ ]:

from sklearn import metrics
y_hat = est.predict(x)
print("MSE: ", metrics.mean_squared_error(y_hat, y))
print("R^2: ", metrics.r2_score(y_hat, y))
print('var: ', y.var())

MSE:  10.539131639803488
R^2:  0.5067870382100248
var:  extent    31.98324
dtype: float64

 

boston = pd.read_excel('/content/drive/MyDrive/Colab Notebooks/BostonHousing.xls')
boston

 

0.00632	18.0	2.31	0	0.538	6.575	65.2	4.0900	1	296	15.3	396.90	4.98	24.0	0
0.02731	0.0	7.07	0	0.469	6.421	78.9	4.9671	2	242	17.8	396.90	9.14	21.6	0
0.02729	0.0	7.07	0	0.469	7.185	61.1	4.9671	2	242	17.8	392.83	4.03	34.7	1
0.03237	0.0	2.18	0	0.458	6.998	45.8	6.0622	3	222	18.7	394.63	2.94	33.4	1
0.06905	0.0	2.18	0	0.458	7.147	54.2	6.0622	3	222	18.7	396.90	5.33	36.2	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
0.06263	0.0	11.93	0	0.573	6.593	69.1	2.4786	1	273	21.0	391.99	9.67	22.4	0
0.04527	0.0	11.93	0	0.573	6.120	76.7	2.2875	1	273	21.0	396.90	9.08	20.6	0
0.06076	0.0	11.93	0	0.573	6.976	91.0	2.1675	1	273	21.0	396.90	5.64	23.9	0
0.10959	0.0	11.93	0	0.573	6.794	89.3	2.3889	1	273	21.0	393.45	6.48	22.0	0
0.04741	0.0	11.93	0	0.573	6.030	80.8	2.5050	1	273	21.0	396.90	7.88	11.9	0

506 rows × 15 columns

 

In [ ]:

print('Shape of data: ', boston.shape)
print('Feature names: ', boston.columns)

Shape of data:  (506, 15)
Feature names:  Index(['CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS', 'RAD', 'TAX',
       'PTRATIO', 'B', 'LSTAT', 'MEDV', 'CAT. MEDV'],
      dtype='object')

In [ ]:

from sklearn import datasets
from sklearn.datasets import fetch_california_housing

california_housing = fetch_california_housing()
X_california_housing, y_california_housing = california_housing.data, california_housing.target
print("Shape of data: ", X_california_housing.shape, y_california_housing.shape)
print("Feature names: ", california_housing.feature_names)

Shape of data:  (20640, 8) (20640,)
Feature names:  ['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population', 'AveOccup', 'Latitude', 'Longitude']

from sklearn import linear_model
train_size = int(len(X_california_housing) / 2)
X_train = X_california_housing[:train_size]
X_test = X_california_housing[train_size:]
y_train = y_california_housing[:train_size]
y_test = y_california_housing[train_size:]
print('Training and testing set sizes', X_train.shape, X_test.shape)

regr = LinearRegression()
regr.fit(X_train, y_train)
print('Coeff and intercept: ', regr.coef_, regr.intercept_)
print('Testing Score: ', regr.score(X_test, y_test))
print('Training MSE: ', np.mean((regr.predict(X_train)- y_train)**2))
print('Testing MSE: ', np.mean((regr.predict(X_test)- y_test)**2))

Training and testing set sizes (10320, 8) (10320, 8)
Coeff and intercept:  [ 4.49445675e-01  5.25146897e-03 -1.15187965e-01  6.49831604e-01
 -4.42831175e-06 -7.49431094e-03 -4.12760239e-01 -3.65411827e-01] -28.871818045412486
Testing Score:  0.586194727280809
Training MSE:  0.5427336522220854
Testing MSE:  0.5482126919151005

In [ ]:

regr_lasso = linear_model.Lasso(alpha = .3)
regr_lasso.fit(X_train, y_train)
print('Coeff and intercept: ', regr_lasso.coef_)
print('Testing Score: ', regr_lasso.score(X_test, y_test))
print('Training MSE: ', np.mean((regr_lasso.predict(X_train)- y_train)**2))
print('Testing MSE: ', np.mean((regr_lasso.predict(X_test)- y_test)**2))

Coeff and intercept:  [ 3.39722672e-01  1.15439467e-02  0.00000000e+00  0.00000000e+00
  4.46333418e-05 -1.97394316e-03 -3.33472332e-02  0.00000000e+00]
Testing Score:  0.4767976308497144
Training MSE:  0.6586900574320259
Testing MSE:  0.69314288173142

In [ ]:

ind = np.argsort(np.abs(regr_lasso.coef_))

ind_feature_names = [california_housing.feature_names[i] for i in ind]

print('Ordered variables (from less to more important):', ind_feature_names)

Ordered variables (from less to more important): ['AveRooms', 'AveBedrms', 'Longitude', 'Population', 'AveOccup', 'HouseAge', 'Latitude', 'MedInc']

In [ ]:

import sklearn.feature_selection as fs
selector = fs.SelectKBest(score_func = fs.f_regression, k = 5)
selector.fit_transform(X_train, y_train)
selector.fit(X_train, y_train)
print('Selected features: ', list(zip(selector.get_support(), california_housing.feature_names)))

Selected features:  [(True, 'MedInc'), (True, 'HouseAge'), (True, 'AveRooms'), (False, 'AveBedrms'), (False, 'Population'), (False, 'AveOccup'), (True, 'Latitude'), (True, 'Longitude')]

 

*그래프는 붙여넣기 오류로 파일 참조