CODEWITHSUNDEEP
pythonpandasmatplotlibmachine-learningscikit-learn
user12907213
user12907213

Reputation:

Add a regression line on the plot with actual data

I have the following data in a pandas dataframe:

freq = [10, 2, 1, 10, 6, 4, 1, 1,
 6, 3, 4, 10, 6, 3, 9, 5,
 5, 5, 4, 2, 2, 9, 11, 7, 5,
 1, 3, 10, 7, 5, 5, 5, 8,
 7, 25, 17, 9, 6, 7, 8, 4,
 10, 3, 1, 7, 11, 6, 5, 10,
 11, 8, 11, 15, 4, 6, 11, 6,
 10, 10, 10, 4, 5, 7, 15, 15,
 10, 12, 17, 25, 26, 22, 14, 15,
 15, 7, 9, 8, 6, 1]

date=[737444, 737445, 737446, 737447, 737448,
 737449, 737450, 737451, 737452, 737453, 737454, 737455, 737456,
 737457, 737458, 737459, 737460, 737461, 737462, 737463, 737464,
 737465, 737466, 737467, 737468, 737469, 737470, 737472, 737473,
 737474, 737475, 737476, 737477, 737478, 737479, 737480, 737481,
 737482, 737483, 737484, 737485, 737486, 737487, 737488, 737489, 
 737490, 737491, 737492, 737493, 737494, 737495, 737496, 737497, 
 737498, 737499, 737500, 737501, 737502, 737503, 737504, 737505,
 737506, 737507, 737508, 737509, 737510, 737511, 737512, 737513, 
 737514, 737515, 737516, 737517, 737518, 737519, 737520, 737521,
 737522, 737523]

I have calculated the coefficient and intercept for regression as follows:

    from sklearn.model_selection import train_test_split
    
    y = np.asarray(df['Frequency'])
    X = df[['Date']]
    X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.3,random_state=42)
    
    model = LinearRegression()
    model.fit(X_train, y_train)
    model.score(X_train, y_train) 
    
    coefs = zip(model.coef_, X.columns)
    model.__dict__

Getting the following results:

    Coefficient: 
     [0.08711929]
    Intercept: 
     -64241.58584385233
    sl = -64241.6 + 0.1 Date

I would like to plot this line above the plot that shows the trend of actual data. How can I do?

Upvotes: 3

Views: 865

Answers (4)

yatu
yatu

Reputation: 88236

A linear regression's model is defined by X*model.coef_ + model.intercept_ (essentially, the result of the prediction). So you can plot the slope along with a scatter plot of the data for a nice visualization of the result:

freq=[10, 2, 1, 10, 6, 4, 1, 1, 6, 3, 4, 10, 6, 3, 9, 5, 5, 5, 4, 2, 2, 9, 11, 7, 5, 1, 3, 10, 7, 5, 5, 5, 8, 7, 25, 17, 9, 6, 7, 8, 4, 10, 3, 1, 7, 11, 6, 5, 10, 11, 8, 11, 15, 4, 6, 11, 6, 10, 10, 10, 4, 5, 7, 15, 15, 10, 12, 17, 25, 26, 22, 14, 15, 15, 7, 9, 8, 6, 1]

date=[737444, 737445, 737446, 737447, 737448, 737449, 737450, 737451, 737452, 737453, 737454, 737455, 737456, 737457, 737458, 737459, 737460, 737461, 737462, 737463, 737464, 737465, 737466, 737467, 737468, 737469, 737470, 737472, 737473, 737474, 737475, 737476, 737477, 737478, 737479, 737480, 737481, 737482, 737483, 737484, 737485, 737486, 737487, 737488, 737489, 737490, 737491, 737492, 737493, 737494, 737495, 737496, 737497, 737498, 737499, 737500, 737501, 737502, 737503, 737504, 737505, 737506, 737507, 737508, 737509, 737510, 737511, 737512, 737513, 737514, 737515, 737516, 737517, 737518, 737519, 737520, 737521, 737522, 737523]

X = np.array(date)
y = np.array(freq)

X_train, X_test, y_train, y_test = train_test_split(X[:,None],y,test_size=0.3,
                                                   random_state=42)

model = LinearRegression()
model.fit(X_train, y_train)

plt.subplots(figsize=(15, 8))
plt.scatter(date, freq, color='lightblue')
# here's how using the actual equation so is is clear how it works
# essentially same as model.precict(X)
plt.plot(date, X*model.coef_ + model.intercept_)

enter image description here

Upvotes: 0

Scott Boston
Scott Boston

Reputation: 153460

Use model.predict:

import numpy as  np
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
import pandas as pd
import matplotlib.pyplot as plt
    
y = np.asarray(freq)
X = np.array(date).reshape(-1,1)
X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.3,random_state=42)
    
model = LinearRegression()
model.fit(X_train, y_train)
model.score(X_train, y_train) 
    
y_pred = model.predict(X)
model.__dict__

plt.subplots(figsize=(15, 8))
plt.scatter(date, freq)
plt.plot(date, y_pred)

Output:

enter image description here

Upvotes: 0

Dorian
Dorian

Reputation: 1488

Use matplotlib.pyplot for plotting.

import numpy as np
import matplotlib.pyplot as plt

freq_reg = intersect + slope * np.array(date, np.dtype=float)

fig, ax = plt.subplot(1, 1)
ax.scatter(date, freq)
plt.plot(date, freq_reg)

plt.set_xlabel('date')
plt.set_ylabel('frequency')

plt.show()

Upvotes: 0

Rupesh
Rupesh

Reputation: 538

import matplotlib.pyplot as plt

datemin = min(date)
datemax = max(date)

x_new = np.linspace(datemin, datemax , 100)
y_new = model.predict(x_new[:, np.newaxis])

plt.figure(figsize=(4, 3))
ax = plt.axes()
ax.scatter(date, freq)
ax.plot(x_new, y_new)

ax.set_xlabel('x')
ax.set_ylabel('y')

ax.axis('tight')
plt.show()

Check a detailed explanation here

Upvotes: 1

Related Questions