Scribd
Upload
6 views2 pages

620 Case Study3

The document outlines a case study on using logistic regression to predict flight arrival status based on a dataset of commercial flights from Washington, DC to New York. It includes tasks such as data preprocessing, model development, and comparison with a classification tree model, focusing on the importance of binary variables for logistic regression. Key steps involve exploring the dataset, training the model, making predictions, and analyzing confusion matrices and lift charts to evaluate model performance.

Uploaded by

kiranodedra1999
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views2 pages

620 Case Study3

The document outlines a case study on using logistic regression to predict flight arrival status based on a dataset of commercial flights from Washington, DC to New York. It includes tasks such as data preprocessing, model development, and comparison with a classification tree model, focusing on the importance of binary variables for logistic regression. Key steps involve exploring the dataset, training the model, making predictions, and analyzing confusion matrices and lift charts to evaluate model performance.

Uploaded by

kiranodedra1999
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Case Study #3: Logistic Regression for Flight

Status
The FlightDelays.csv file contains a data set with the information on all commercial flights departing the
Washington, DC area and arriving at one of the New York area airports. For each flight, there is information
on the departure and arrival airports, the distance of the route, the scheduled time and day of the flight,
and so on. The outcome that you need to predict is the flight arrival status, i.e., whether a flight is delayed
or on time (FL_STATUS). A delay is defined as an arrival that is at least 15 minutes later than scheduled.
The table below descries each of the columns in the data set.

Variables Description of Variables


DEP_TIME Scheduled departure time.
CARRIER Carrier abbreviation: CO=Continental, DH=Discovery Airways,
DL= Delta Airlines, MQ=American Eagle, OH=PSA Airlines,
RU=AirBridge Cargo, UA=United Airlines, US=US Airways.
DEP_TIME Actual flight departure time.
DEST Abbreviation of the destination airport in New York area:
EWR=Newark, JFK=John F. Kennedy, LGA= LaGuardia.
DISTANCE Distance of the route in miles.
FL_NUM Flight number.
ORIGIN Abbreviation of the DC area airport: BWI=Baltimore
/Washington, DCA=Ronald Reagan Washington, and IAD=Dulles
International.
WEATHER Weather condition for a flight: good flying condition = 0, poor
flying condition = 1.
WK_ DAY Day of the week, from Monday = 1 through Sunday = 7.
MTH_DAY Day of the month, from 1 through 31.
FL_STATUS Flight arrival status, two classes: ‘delayed’ and ‘ontime’.

Questions

1. Upload, explore, clean, and preprocess data.


a. Why a logistic regression model may be used in this case? Why may you not apply a
multiple linear regression model in this case? Provide brief answers to both questions.
b. Create a flight_df data frame by uploading the original data set into Python. Remove
‘DEST’ and ‘ORIGIN’ variables from the flight_df data frame. Convert ‘CARRIER’ and
‘FL_STATUS’ into binary variables. This question 1b of part 1 will not be graded.
c. Why does the output variable ‘FL_STATUS’ need to be converted into binary variables for
logistic regression? Briefly explain.

2. Develop a logistic regression model for the Flight Delays case.


a. Develop in Python the predictor variables (14 variables) and outcome variable
(‘FL_STATUS’) and partition the data set (80% for training and 20% for validation
partitions, random_state=1). Train a logistic regression model using LogisticRegression()
with the training data set and display in Python the model’s parameters (intercept and
regression coefficients). Provide these parameters in your report and also present the
mathematical equation of the trained logistic regression model.
b. In Python, make predictions and identify probabilities p(0) and p(1) for the validation data
set. For the first 20 records in the validation data set, display a table that contains the
actual and predicted flight arrival status, and probabilities p(0) and p(1). Present this table
in your report, and comment on the predicted vs. actual flight arrival status.
c. Identify and display in Python confusion matrices for the training and validation
partitions. Present them in your report and comment on accuracy (misclassification) rate
for both partitions and explain if there is a possibility of overfitting.
d. Create and display in Python the Lift chart only for ‘delayed’ flight status. For that, use
p(0) for .sort_values() and p(0) in liftChart(). Also use ncols=1 in plt.subplots() for a single
plot, and remove ax=axes[1] from liftChart(). Present this Lift chart in your report and
briefly explain what the chart demonstrates and what conclusion(s) can be made.

3. Compare results of logistic regression model vs. classification tree model for the same data set.
a. Present and compare in your report the validation confusion matrix for the logistic
regression model in 2c of this case versus the validation confusion matrix using the
GridSearchCV() algorithm for the classification tree in the previous case study #2. Using
the accuracy value (misclassification rate), which model would you recommend applying
for classification (prediction) of flight arrival status? Briefly explain your answer.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy