Initial linear regression

One . use Excel In the data analysis function, do linear regression

1. take 20 Group data

2. take 200 Group data

3. take 2000 Group data

 

 

Two . use jupyter Programming （ Without the help of third-party libraries ）, Linear regression with least square method

1.python Least squares source code

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
points = np.genfromtxt("D:/wh.csv",delimiter=",")
# take wh.csv The data in the file is assigned to points
# take points The data in are assigned to x,y, Find regression equation y=ax+b
x=points[1:21,1];
y=points[1:21,2];
pccs = np.corrcoef(x, y)
c,d=pccs
e,f=c
x_mean = np.mean(x)
y_mean = np.mean(y)
xsize = x.size
zi = (x * y).sum() - xsize * x_mean *y_mean
mu = (x ** 2).sum() - xsize * x_mean ** 2
a = zi / mu
b = y_mean - a * x_mean
a = np.around(a,decimals=2)
b = np.around(b,decimals=2)
print(f' Regression line equation :y = {a}x + {b}')
print(f' The correlation coefficient is 0 {f}')
# Using third party libraries skleran Draw a fit curve 
y1 = a*x + b
plt.scatter(x,y)
plt.plot(x,y1,c='r')

2. take 20 Group data

 

 

3. take 200 Group data

 

 

 

 

 

4. take 2000 Group data

  3、 ... and . use jupyter Programming , With the help of skleran Do linear regression

 1.python With the help of skleran Source code

from sklearn import linear_model        # Express , You can call sklearn Medium linear_model The module performs linear regression .
import matplotlib.pyplot as plt
import numpy as np
%matplotlib inline
data = np.genfromtxt("wh.csv",delimiter=",")
data1=data[1:21]# Change the value according to the data taken 
x=[example[1] for example in data1]
y=[example[2] for example in data1]
pccs = np.corrcoef(x, y)
c,d=pccs
e,f=c
X = np.asarray(x).reshape(-1, 1)
Y = np.asarray(y).reshape(-1, 1)
model = linear_model.LinearRegression()
model.fit(X,Y)
b=model.intercept_[0] # intercept 
a=model.coef_[0]# Coefficient of linear model 
a1=a[0]
print(f' Regression line equation :y = {a1}x + {b}')
print(f' The correlation coefficient is 0 {f}')
y1 = a1*X + b
plt.scatter(X,Y)
plt.plot(x,y1,c='r')

2. take 20 Group data

 

3. take 200 Group data

 

 

 

4. take 2000 Group data

 

  Compare ： It can be seen from the three ways that the regression equation is generally the same , There are some differences in the correlation coefficient

reference ：(3 Bar message ) Separate use Excel,jupyter,jupyter+skleran Do linear regression exercises _ Parallel leaf blog -CSDN Blog