kangeugine

import numpy as np

def ols(x,y):
    x_mean = x.mean()
    y_mean = y.mean()
    beta_1 = np.multiply((x - x_mean), (y - y_mean)).sum() / np.square((x - x_mean)).sum()
    beta_0 = y_mean - (beta_1 * x_mean)

    def lin_model(new_x, slope=beta_1, intercept=beta_0):
        return new_x * slope + intercept

kangeugine

def ols(x,y):
    x_mean = x.mean()
    y_mean = y.mean()
    beta_1 = np.multiply((x - x_mean), (y - y_mean)).sum() / np.square((x - x_mean)).sum()
    beta_0 = y_mean - (beta_1 * x_mean)
    
    def lin_model(new_x, slope=beta_1, intercept=beta_0):
        return new_x * slope + intercept

    return lin_model

kangeugine

def is_palindrome(sequence):
	return sequence == sequence[::-1]

# test when sequence IS palindrome
def test_true():
	seq = 'bob'
	assert is_palindrome(seq) == True

def test_true_list():
	seq = [1, 0, 1]

kangeugine

import pytest

@pytest.fixture
def default_list():
	return [1, 2]

def test_fixture_1(default_list):
	del default_list[0]
	assert default_list == [2]

kangeugine

import pytest

def is_int(age):
	return isinstance(age, int)

@pytest.mark.parametrize("age", [
	"five",
	1.5,
	{2: 3},
	[2, 5],

kangeugine

import numpy as np

def ols(x,y):
    x_mean = x.mean()
    y_mean = y.mean()
    beta_1 = np.multiply((x - x_mean), (y - y_mean)).sum() / np.square((x - x_mean)).sum()
    beta_0 = y_mean - (beta_1 * x_mean)
    
    def lin_model(new_x, slope=beta_1, intercept=beta_0):
        return new_x * slope + intercept

kangeugine

from scipy import optimize
import matplotlib.pyplot as plt
import numpy as np

x = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ,11, 12, 13, 14, 15], dtype=float)
y = np.array([5, 7, 9, 11, 13, 15, 28.92, 42.81, 56.7, 70.59, 84.47, 90, 80, 70, 60])

def piecewise_linear(x, y0, y1, b0, b1, b2):
    x0 = 6
    x1 = 12

kangeugine

import pandas as pd
import matplotlib.pyplot as plt
from datetime import datetime

def parser(x):
    return datetime.strptime('190'+x, '%Y-%m')
  
series = pd.read_csv('data/sales-of-shampoo-over-a-three-ye.csv', \
                     header=0, \
                     parse_dates=[0], \

kangeugine

from pandas.tools.plotting import autocorrelation_plot

autocorrelation_plot(series)
plt.show()

kangeugine

from statsmodels.tsa.arima_model import ARIMA

# fit model
model = ARIMA(series, order=(5,1,0))
model_fit = model.fit(disp=0)
# print(model_fit.summary())

# plot residual erros
residuals = pd.DataFrame(model_fit.resid)
residuals.plot()