Machine Learning with Scikit-Learn 1 - Fundamentals of ML

Load and prepare Life satisfaction data

import pandas as pd

 # Download CSV from http://stats.oecd.org/index.aspx?DataSetCode=BLI

oecd_bli = pd.read_csv(datapath+"oecd_bli_2015.csv", thousands=',') 

oecd_bli = oecd_bli[oecd_bli["INEQUALITY"]=="TOT"] 

oecd_bli = oecd_bli.pivot(index="Country", columns="Indicator", values="Value") 

oecd_bli.head(2)
oecd_bli["Life satisfaction"].head()

Load and prepare GDP per capita data

# Download data from http://goo.gl/j1MSKe (=> imf.org)

gdp_per_capita = pd.read_csv(datapath+"gdp_per_capita.csv", thousands=',', delimiter='\t', encoding='latin1', na_values="n/a") 

gdp_per_capita.rename(columns={"2015": "GDP per capita"}, inplace=True) gdp_per_capita.set_index("Country", inplace=True) 

gdp_per_capita.head(2)

full_country_stats = pd.merge(left=oecd_bli, right=gdp_per_capita, left_index=True, right_index=True)

full_country_stats.sort_values(by="GDP per capita", inplace="True")

full_country_stats

remove_indices = [0, 1, 6, 8, 33, 34, 35] 

keep_indices = list(set(range(36)) - set(remove_indices)) 

sample_data = full_country_stats[["GDP per capita", 'Life satisfaction']].iloc[keep_indices] missing_data = full_country_stats[["GDP per capita", 'Life satisfaction']].iloc[remove_indices]

sample_data.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(5,3))

plt.axis([0, 60000, 0, 10]) 

position_text = { "Hungary": (5000, 1), 

 "Korea": (18000, 1.7), 

 "France": (29000, 2.4), 

 "Australia": (40000, 3.0), 

 "United States": (52000, 3.8), } 

for country, pos_text in position_text.items(): 

pos_data_x, pos_data_y = sample_data.loc[country] 

country = "U.S." if country == "United States" else country 

plt.annotate(country, xy=(pos_data_x, pos_data_y), xytext=pos_text, arrowprops=dict(facecolor='black', width=0.5, shrink=0.1, headwidth=5)) 

plt.plot(pos_data_x, pos_data_y, "ro") 

save_fig('money_happy_scatterplot') 

plt.show()

import numpy as np

sample_data.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(5,3)) 

plt.axis([0, 60000, 0, 10]) 

X=np.linspace(0, 60000, 1000) 

plt.plot(X, 2*X/100000, "r") 

plt.text(40000, 2.7, r"$\theta_0 = 0$", fontsize=14, color="r") 

plt.text(40000, 1.8, r"$\theta_1 = 2 \times 10^{-5}$", fontsize=14, color="r") 

plt.plot(X, 8 - 5*X/100000, "g") 

plt.text(5000, 9.1, r"$\theta_0 = 8$", fontsize=14, color="g") 

plt.text(5000, 8.2, r"$\theta_1 = -5 \times 10^{-5}$", fontsize=14, color="g") 

plt.plot(X, 4 + 5*X/100000, "b") 

plt.text(5000, 3.5, r"$\theta_0 = 4$", fontsize=14, color="b") 

plt.text(5000, 2.6, r"$\theta_1 = 5 \times 10^{-5}$", fontsize=14, color="b") save_fig('tweaking_model_params_plot')

plt.show()

from sklearn import linear_model

lin1 = linear_model.LinearRegression() 

Xsample = np.c_[sample_data["GDP per capita"]]

ysample = np.c_[sample_data["Life satisfaction"]] 

lin1.fit(Xsample, ysample) 

t0, t1 = lin1.intercept_[0], lin1.coef_[0][0] 

t0, t1

sample_data.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(5,3))

plt.axis([0, 60000, 0, 10]) 

X=np.linspace(0, 60000, 1000) 

plt.plot(X, t0 + t1*X, "b") 

plt.text(5000, 3.1, r"$\theta_0 = 4.85$", fontsize=14, color="b") 

plt.text(5000, 2.2, r"$\theta_1 = 4.91 \times 10^{-5}$", fontsize=14, color="b") save_fig('best_fit_model_plot') 

plt.show()

cyprus_gdp_per_capita = gdp_per_capita.loc["Cyprus"]["GDP per capita"] print(cyprus_gdp_per_capita)

cyprus_predicted_life_satisfaction = lin1.predict(cyprus_gdp_per_capita)[0][0]

cyprus_predicted_life_satisfaction

sample_data.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(5,3), s=1) X=np.linspace(0, 60000, 1000)

plt.plot(X, t0 + t1*X, "b") 

plt.axis([0, 60000, 0, 10]) 

plt.text(5000, 7.5, r"$\theta_0 = 4.85$", fontsize=14, color="b") 

plt.text(5000, 6.6, r"$\theta_1 = 4.91 \times 10^{-5}$", fontsize=14, color="b") plt.plot([cyprus_gdp_per_capita, cyprus_gdp_per_capita], [0, cyprus_predicted_life_satisfaction], "r--") 

plt.text(25000, 5.0, r"Prediction = 5.96", fontsize=14, color="b") 

plt.plot(cyprus_gdp_per_capita, cyprus_predicted_life_satisfaction, "ro") save_fig('cyprus_prediction_plot') 

plt.show()

sample_data[7:10]
(5.1+5.7+6.5)/3

backup = oecd_bli, gdp_per_capita

 def prepare_country_stats(oecd_bli, gdp_per_capita): 

 return sample_data 

 # Code example ########################################################################

import sklearn 

import numpy as np 

import pandas as pd 

 # Load the data 

oecd_bli = pd.read_csv(datapath+"oecd_bli_2015.csv", thousands=',') 

gdp_per_capita = pd.read_csv(datapath+"gdp_per_capita.csv", thousands=',',delimiter='\t', encoding='latin1', na_values="n/a") 

 # Prepare the data 

country_stats = prepare_country_stats(oecd_bli, gdp_per_capita) 

X = np.c_[country_stats["GDP per capita"]] 

y = np.c_[country_stats["Life satisfaction"]] 

# Visualize the data country_stats.plot(kind='scatter', x="GDP per capita", y='Life satisfaction') plt.show() 

 # Select a linear model 

lin_reg_model = sklearn.linear_model.LinearRegression() 

# Train the model 

lin_reg_model.fit(X, y) 

 # Make a prediction for Cyprus 

X_new = [[22587]] # Cyprus' GDP per capita 

print(lin_reg_model.predict(X_new)) # outputs [[ 5.96242338]]
######################################################################## 

 oecd_bli, gdp_per_capita = backup

position_text2 = {

 "Brazil": (1000, 9.0), 

 "Mexico": (11000, 9.0), 

 "Chile": (25000, 9.0), 

 "Czech Republic": (35000, 9.0), 

 "Norway": (60000, 3), 

 "Switzerland": (72000, 3.0), 

 "Luxembourg": (90000, 3.0), }

sample_data.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(8,3)) 

plt.axis([0, 110000, 0, 10]) 

for country, pos_text in position_text2.items():

pos_data_x, pos_data_y = missing_data.loc[country] 

plt.annotate(country, xy=(pos_data_x, pos_data_y), xytext=pos_text, arrowprops=dict(facecolor='black', width=0.5, shrink=0.1, headwidth=5)) 

plt.plot(pos_data_x, pos_data_y, "rs") 

X=np.linspace(0, 110000, 1000) 

plt.plot(X, t0 + t1*X, "b:") 

lin_reg_full = linear_model.LinearRegression() 

Xfull = np.c_[full_country_stats["GDP per capita"]] 

yfull = np.c_[full_country_stats["Life satisfaction"]] 

lin_reg_full.fit(Xfull, yfull) 
t0full, t1full = lin_reg_full.intercept_[0], lin_reg_full.coef_[0][0] 

X = np.linspace(0, 110000, 1000) 

plt.plot(X, t0full + t1full * X, "k") 

save_fig('representative_training_data_scatterplot') 

plt.show()

full_country_stats.plot(kind='scatter', x="GDP per capita", y='Life satisfaction', figsize=(8,3)) plt.axis([0, 110000, 0, 10])

from sklearn import preprocessing
from sklearn import pipeline
poly = preprocessing.PolynomialFeatures(degree=60, include_bias=False)
scaler = preprocessing.StandardScaler()
lin_reg2 = linear_model.LinearRegression()
pipeline_reg = pipeline.Pipeline([('poly', poly), ('scal', scaler), ('lin', lin_reg2)])
pipeline_reg.fit(Xfull, full)
curve = pipeline_reg.predict(X[:, np.newaxis])
plt.plot(X, curve)
save_fig('overfitting_model_plot')
plt.show()

full_country_stats.loc[[c for c in full_country_stats.index if "W" in c.upper()]]["Life satisfaction"]

gdp_per_capita.loc[[c for c in gdp_per_capita.index if "W" in c.upper()]].head()

plt.figure(figsize=(8,3))

plt.xlabel("GDP per capita") 

plt.ylabel('Life satisfaction') 

plt.plot(list(sample_data["GDP per capita"]), list(sample_data["Life satisfaction"]), "bo") plt.plot(list(missing_data["GDP per capita"]), list(missing_data["Life satisfaction"]), "rs") 

X = np.linspace(0, 110000, 1000) 

plt.plot(X, t0full + t1full * X, "r--", label="Linear model on all data") 

plt.plot(X, t0 + t1*X, "b:", label="Linear model on partial data") 

ridge = linear_model.Ridge(alpha=10**9.5) 

Xsample = np.c_[sample_data["GDP per capita"]] 

ysample = np.c_[sample_data["Life satisfaction"]] 

ridge.fit(Xsample, sample) 

t0ridge, t1ridge = ridge.intercept_[0], ridge.coef_[0][0] 

plt.plot(X, t0ridge + t1ridge * X, "b", label="Regularized linear model on partial data") plt.legend(loc="lower right") 

plt.axis([0, 110000, 0, 10]) 

plt.show()

backup = oecd_bli, gdp_per_capita def

prepare_country_stats(oecd_bli, gdp_per_capita): 

return sample_data 

# Code example ######################################################################## 

from sklearn import neighbors 

import numpy as np 

import pandas as pd 

# Load the data 

oecd_bli = pd.read_csv(datapath+"oecd_bli_2015.csv", thousands=',') 

gdp_per_capita = pd.read_csv(datapath+"gdp_per_capita.csv", thousands=',',delimiter='\t', encoding='latin1', na_values="n/a") 

# Prepare the data 

country_stats = prepare_country_stats(oecd_bli, gdp_per_capita) 

X = np.c_[country_stats["GDP per capita"]] 

y = np.c_[country_stats["Life satisfaction"]] 

# Visualize the data 

country_stats.plot(kind='scatter', x="GDP per capita", y='Life satisfaction') 

plt.show() 

# Select a k-neighboors regression model 

k_neigh_reg_model = neighbors.KNeighborsRegressor(n_neighbors=3) 

# Train the model 

k_neigh_reg_model.fit(X, y) 

# Make a prediction for Cyprus 

X_new = [[22587]] 

# Cyprus' GDP per capita 

print(lin_reg_model.predict(X_new)) 

######################################################################## 

oecd_bli, gdp_per_capita = backup

https://github.com/ageron/handson-ml/blob/master/01_the_machine_learning_landscape.ipynb