Scribd
Upload
29 views5 pages

Ekonometrika

The document discusses assessing assumptions for linear regression models, including independence of residuals, homoscedasticity (constant variance of residuals), and lack of autocorrelation. Graphs and statistical tests are used to evaluate these assumptions. For the sample model, the residuals appear randomly scattered with no discernible pattern over time (independence), have consistent spread in the residuals vs fitted values plot (homoscedasticity), and the Durbin-Watson and Breusch-Pagan tests do not show significant autocorrelation or heteroscedasticity, indicating the assumptions are reasonable for the model.

Uploaded by

Qeti Macharashvili
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
29 views5 pages

Ekonometrika

The document discusses assessing assumptions for linear regression models, including independence of residuals, homoscedasticity (constant variance of residuals), and lack of autocorrelation. Graphs and statistical tests are used to evaluate these assumptions. For the sample model, the residuals appear randomly scattered with no discernible pattern over time (independence), have consistent spread in the residuals vs fitted values plot (homoscedasticity), and the Durbin-Watson and Breusch-Pagan tests do not show significant autocorrelation or heteroscedasticity, indicating the assumptions are reasonable for the model.

Uploaded by

Qeti Macharashvili
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

2nd slide …….

3rd- 4th slide

The summary() function provides a quick overview of key statistics for each variable in the dataset.

The summary includes measures such as the minimum, 1st quartile, median, mean, 3rd quartile, and
maximum for the Murder variable.

The str() function provides information about the structure of the dataset, including variable names,
types, and the first few observations.

The structure shows the data type of each variable (e.g., numeric) and the first few observations. It
helps you understand the type of data you're working with and the potential need for data
preprocessing.

შეიძლება ამოღება :

Overall Conclusions:

Data Range: You can observe the range of values for each variable, understanding the spread and
distribution.

Data Types: Identifying whether variables are numeric or categorical is crucial for selecting
appropriate analysis methods.

Potential Outliers: Extreme values or outliers may be evident from summary statistics.

Data Size: The number of observations (rows) is visible, providing an idea of the dataset's size.
It seems like your question is about descriptive statistics and measures of central tendency and
variability for a variable. Let me explain each of these measures:

1. **Minimum:** The smallest value in a dataset.

2. **1st Quartile (Q1):** The value below which 25% of the data falls. It is also known as the lower
quartile.

3. **Median (or 2nd Quartile, Q2):** The middle value of a dataset when it is ordered. If there is an
even number of observations, the median is the average of the two middle values.

4. **Mean:** The average of all values in the dataset. It is calculated by summing up all values and
dividing by the number of observations.

5. **3rd Quartile (Q3):** The value below which 75% of the data falls. It is also known as the upper
quartile.

6. **Maximum:** The largest value in a dataset.

To find these measures, you need a dataset. If you have a set of data, you can organize it in
ascending order and then calculate each measure accordingly. If you provide a dataset, I can help
you compute these statistics.

14th slide
1.1 Independence of Residuals 14, 15, 16, 17(ქეთი)
Time-series plot of residuals

Code –

Otput –
In the "Time-Series Plot of Residuals":

In the Ideal Scenario, The plot should not show any discernible pattern or trend over time. The
residuals should appear randomly scattered without any systematic structure. If you observe a
clear pattern, oscillation, or any structure, it might indicate a lack of independence among
residuals, violating the assumption.

In out case, plot does not show any pattern or trend over time, so the independence of residuals
assumption is reasonable for the model

15th slide

Durbin-Watson test (Tests for absence of autocorrelation of regression residuals)

Code/output

The Durbin-Watson test is a statistical test that formally evaluates the presence of
autocorrelation (dependence between residuals) in the residuals. In the Ideal Scenario,The
Durbin-Watson test statistic should be close to 2. A value around 2 suggests no significant
autocorrelation, indicating independence of residuals. If the test statistic is significantly different
from 2, it may indicate autocorrelation. A value less than 2 suggests positive autocorrelation,
while a value greater than 2 suggests negative autocorrelation.

In our case, the value is 1.7, it is close to 2, therefore, the independence of residuals assumption
is reasonable for the model.

16th slide

1.1 Homoscedasticity (Constant Variance of Residuals):

Scatter plot of residuals against predicted values

Code -

Output –

In the context of the "Scale-Location Plot" and the Breusch-Pagan test, we are assessing the
homoscedasticity (constant variance of residuals) assumption in the linear regression model.

In the "Scale-Location Plot" (also known as the spread-location plot) in Ideal Scenario the
plot should show a random scatter of points with no clear pattern. If you observe any
systematic change in the spread of residuals, it might indicate a violation of the
homoscedasticity assumption.

In our case, the "Scale-Location Plot" appears to have a consistent spread of points without
any clear pattern, hence the homoscedasticity assumption is reasonable for our model.
17th slide
Breusch-Pagan test

Code/output

The Breusch-Pagan test is a statistical test that formally checks for heteroscedasticity (non-
constant variance of residuals) in the model. In the Ideal Scenario A non-significant result
(high p-value) suggests that the assumption of homoscedasticity is reasonable. If the p-value
is low (below a significance threshold, e.g., 0.05), it suggests evidence of heteroscedasticity,
indicating a potential violation of the assumption.

In our case, p – value is 0.1294, so the homoscedasticity assumption is reasonable for our
model.

You might also like