Scribd
Upload
37 views

Regression Test Lesson Notes (Optional Download)

The document discusses linear regression, describing dependent and independent variables, regression lines, and how to perform simple and multiple linear regression in Python using Scikit-learn. It provides an example using monthly spending and sales data to build a linear regression model and calculate the coefficient of determination.

Uploaded by

Chuyên Mai Tất
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
37 views

Regression Test Lesson Notes (Optional Download)

The document discusses linear regression, describing dependent and independent variables, regression lines, and how to perform simple and multiple linear regression in Python using Scikit-learn. It provides an example using monthly spending and sales data to build a linear regression model and calculate the coefficient of determination.

Uploaded by

Chuyên Mai Tất
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Learning Objective

In this module, you will learn about linear regression and how to perform
regression given a data.

Learning Objective:

1. Perform Regression Test


Regression Test

Regression analysis is a set of statistical processes for estimating the


relationships between a dependent variable and independent variables.

Dependent variable:(DV) is what is impacted by the independent variable.


To clarify, it reflects the change made by independent variable.

Independent variable: (IV) is what we expect will influence of variable. In


other words, this represent the variable that changes on its own.In other
words, it is independent of other variables in the study.

Regression: a statistical technique for finding the best-fitting straight line


to describe the variable relationships for a set of data.

Regression line: the resulting best-fitting straight line.

Regression equation: the resulting best-fitting linear equation.


Regression

Linear Regression

Simple and Multiple Linear Regression

If you determine that your data sets are associated or correlated with each
other, you can perform linear regression on them. Doing so enables you to
predict additional data based on that regression model. There is single
linear when you have one independent variable (x) and one dependent
variable (y). And there is multiple linear regression when you have
multiple independent variables (x's) but only one dependent variable (y).
You can think of a dependent variable as an event that is influenced by the
independent variable. For example, when comparing how much sunlight a
plant receives versus its growth in length, it is probably better to say that
the plant’s growth is more likely dependent on the amount of sunlight than
the reverse. Thus, the plant’s length is the dependent variable and the
amount of sunlight is the independent variable.

When thinking of Linear regression:


* The independent variable is denoted as x.
* The dependent variable, is denoted as y.

Linear Regression finds out a linear relationship between x (input) and


y(output). If we put all of our x values, the linear regression line will give us
the line of best fit.

regre
Lets talk about the graph above. So the blue dots are our actual data points
and the red light is the line of best fit for said graph. A linear regression
model assumes that the relationship is linear. So if we are given an y we
can use our linear regression model to predict the y. The difference is
between our actual and the value on the line at that point is error term.

We will be using sklearn. To help us with the linear regression model in


python.

For example, if you have the following data:

# create our data


month = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
spend =[1000, 4000, 5000, 4500, 3000, 4000, 9000, 11000, 15000,
12000, 7000, 3000]
sales =[9914, 40487, 54324, 50044, 34719, 42551, 94871, 118914,
158484, 131348, 78504, 36284]

# the lists, with columns specified


df = pd.DataFrame(list(zip(month, spend,sales)),
columns =['month', 'spend','sales'])

We are going to perform a simple linear regression where we have spend


and sales as our IV and DV. Lets label them.

# Defining our independent and dependent variables


y=np.array(df['spend'])
x=np.array(df['sales']).reshape(-1, 1)

Now that we have our x and our y. We are going to create our model and
then find the coefficient of determination. The coefficient of determination
or R-squared examines how the difference of the two variables compare to
each other. R-square is a value between 0.0 and 1.0. Where 1.0 represent a
perfect fit. The closer to 0.0 the value is the less reliable we can consider
our model to be.

# create our regression model


model = LinearRegression().fit(x,y)
# get our r_squared
r_sq = model.score(x, y)
print('coefficient of determination:', r_sq)

Now that we know our coefficient of determination, is close to 1.0. We can


make predictions,that will be somewhat close to the actual value. Lets test
it out
Feel free to replace the 39000 with other number to check how close or
accurate it is.

# Here can use the .predict and give it an x value to


print(model.predict(np.array([39000]).reshape(-1, 1)))

Logistic Regression
Logistic regression is performed when you have categorical data in the
mix. For example, if you want to determine if sex is dependent on the
height of person, you should use logistic regression.

You might also like