Scribd
Upload
27 views

Lab 6

This document contains details about Lab Experiment 6.1 on regression and error analysis conducted as part of the Data Mining and Predictive Modelling course. The experiment involves: 1. Creating regression models using variables from a car dataset and analyzing model statistics and diagnostics. 2. Cleaning an auto MPG dataset, selecting relevant columns, dividing data into train and test sets, and using linear regression to train and evaluate a model predicting acceleration from horsepower. 3. Generating actual and predicted values, calculating accuracy, F1-score and precision, creating a confusion matrix, plotting error and ROC curves using true/false positive/negative values.

Uploaded by

Tushar Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
27 views

Lab 6

This document contains details about Lab Experiment 6.1 on regression and error analysis conducted as part of the Data Mining and Predictive Modelling course. The experiment involves: 1. Creating regression models using variables from a car dataset and analyzing model statistics and diagnostics. 2. Cleaning an auto MPG dataset, selecting relevant columns, dividing data into train and test sets, and using linear regression to train and evaluate a model predicting acceleration from horsepower. 3. Generating actual and predicted values, calculating accuracy, F1-score and precision, creating a confusion matrix, plotting error and ROC curves using true/false positive/negative values.

Uploaded by

Tushar Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

School of Computer Science Engineering and Technology

Course- BTech Type- Specialization Core


Course Code- CSET228 Course Name- Data Mining and Predictive
Modelling
Year- 2022-23 Semester- Even
Date- Batch-

Lab # No. (6.1)


Regression and error analysis
CO Mapping
Exp No. Name CO1 CO2 CO3
6.1 Regression and error anlysis 🗸

1. explore regression analysis. (40 Minutes)

a. Create a regression model using ‘citympg’ as the dependent variable and ‘horsepower’ as the
independent variable. Report all model statistics and comment on the model diagnostics.
b. Create the same regression model keeping ‘highwaympg’ as the dependent variable and ‘horsepower’
as the independent variable. Report all model statistics and comment on the model diagnostics.
c. Include a basic scatterplot of the data for both above as well.
d. Create another regression model using ‘citympg’ and ‘highwaympg’ as independent variables
separately and ‘price’ as the dependent variable. Examine in your both regression models whether these
data are positively or negatively correlated or not.
e. Create a regression model using two other variables from the ‘CarPrice_Assignment.csv’ dataset (i.e.,
not ‘highwaympg’, ‘citympg’ or ‘horsepower’ as independent variable). You can keep ‘price’ as
dependent variable. Take any other two separately as independent variable and create two separate
regression model if it can be ensured the existence of a reasonable underlying connection between the
independent and dependent variables.
f. Report all model statistics and comment on the model diagnostics. (20 Min)

2. Use auto-mpg dataset, attribute information of dataset is given in auto-mpg.txt file.


a) Clean dataset i.e., adjust the missing values. Select cylinder, horsepower, mpg and acceleration columns
and create a new dataset.
b) Plot a diagram between (mpg, acceleration) and (cylinder and acceleration). Create a train and test
dataset by dividing the data in 80% and 20% respectively.
c) From the train dataset, plot (horsepower, acceleration), use LinearRegression() from linear_model in
sklearn library to train the model. For training take “Horsepower” as train_x and “Acceleration” as
train_y and finally evaluate the model using testing dataset. . (20 Min)

3. Create two columns using random function. First as actual values and second as predicted values.
School of Computer Science Engineering and Technology
a) On these values find accuracy, f1-score and precision respectively.
b) Create a confusion matrix and display the confusion matrix. Display the error graph between actual and
predicted values.
c) Using confusion matrix find True positive, False positive, True Negative and False Negative. Using these
values plot a ROC curve. . (20 Min)

You might also like