import pandas as pd
import numpy as np
from sklearn.metrics import mean_squared_error
from sklearn.linear_model import LinearRegression
from src.config import config
import sys
import matplotlib.pyplot as plt
def z_score(df):
return (df - np.mean(df))/np.std(df)
if __name__ == '__main__':
PATH = config.get_dir()
country = config.get_country()
adm = config.get_headers(country, 'adm')
cdr = config.get_headers(country, 'cdr')
dhs = config.get_headers(country, 'dhs')
data = pd.DataFrame(pd.read_csv(PATH+'/final/%s/master_2.0.csv'%country,
usecols=['Pagerank', 'G_residuals',
'EigenvectorCentrality',
'BloodPosRateSL', 'Log_pop_density', 'BloodPosRate'])).dropna()
data = data[data['BloodPosRate'] > 0]
data = data.ix[1:]
z_data = pd.DataFrame(z_score(data)).as_matrix()
mse_all = []
for i in range(5):
mse_1, mse_2, mse_3, mse_4, mse_5 = [], [], [], [], []
print 'Training set %d' % i
prop = np.floor(len(data) / 2) + (i * 7)
def get_country_data(country):
PATH = config.get_dir()
return pd.DataFrame(pd.read_csv(PATH+'/final/%s/master_SL.csv' % country))