代写STAT 380、代做DataSummary留学生、R编程调试、代写R

日期：2019-03-01 09:10

STAT 380 Midterm

Spring 2019

YOUR NAME

Due: 03/01/2019 11:59PM

Integrity statment. The work submitted for this assessment is entirely my own. I have neither given or

recieved unauthorized assistance during the assessment, and/or I will speak with Dr. Beckman privately if I

am now aware or later become aware of any activity that may be in violation of academic integrity policies

stated in the course syllabus or Penn State policy. I understand that the content of my completed exam

may not be shared, used, or reproduced for any reason without expressed written permission of Dr. Matthew

Beckman (mdb268@psu.edu).

Front matter

# always clean up R environment

rm(list = ls())

# set RNG seed (don't change this for the midterm)

set.seed(822)

0.1 Introduction

This assessment will include exploration of flights data, which can be accessed from the nycflights13 R

package, as well as weather data provided by the United States National Climactic Data Center (NCDC)

provided to you in the GitHub Classroom repository provided to accompany the take-home portion of the

midterm exam.

“nycWeather2013.csv” contains 2013 daily weather data for three airports in the New York City area.

“nycWeather2013_DataSummary.pdf” is the data description document provided by the NCDC to

accompany our specific data set.

0.2 Logistics

Logistics

Canvas: Submit your completed HTML R Notebook (with embedded Rmd) to Canvas before the due date.

If you are unable to render the R Notebook, you may submit your Rmd directly for partial credit.

GitHub: You are required to make at least 4 substantive GitHub commits during your work on the midterm.

Also, your completed assignment must be entirely reproducible, meaning a grader with access only to your

GitHub Repo should be able to execute your Rmd document and produce your final HTML R Notebook

with absolutely no modification required.

Grading

Overall

– [5 pts] STAT 380 style guide use

1

– [5 pts] GitHub commits

– [5 pts] Reproducibility

– [5 pts] Overall quality

Part 1: Principal Components Analysis of JFK Weather Station

– [10 pts] Data preparation

– [15 pts] PCA of JFK weather station

Part 2: Predictive Modeling of LaGuardia Flight Delays

– [12 pts] Data Preparation

– [4 pts] Partition Test and Training Data

– [12 pts] Predictive modeling

– [10 pts] Visualization

Part 1: Principal Components Analysis of JFK Weather Station

1.1 Data preparation

Task 1.1.1 The data are “complete”, but have been coded in an unconventional way. Recode

all use of NA as 0 in the NCDC data.

Task 1.1.2 A few of the variables in the provided data cannot be used for any meaningful

analysis because every entry is identical. Identify which variables they are, report a bullet list

indicating the name and description of these variables, and then remove them from your data

prior to analysis.

1.2 Principal components analysis for JFK weather station

Task 1.2.1 Perform principal components analysis on data from the JFK airport weather

station (only). Standardize all variables prior to analysis.

Task 1.2.2 Plot the proportion of variance explained by each principal component on a scree

plot.

Task 1.2.3 How many principal components would be required to explain at least 76% of the

variability in the JFK weather station data. Be sure to show your work.

Part 2: Predictive Modeling LaGuardia Flight Delays

2.1 Data preparation

Create LaGuardiaFlights data according to the following instructions:

1. Subset the flights data to include only the carriers that had at least 1000 flights depart from LaGuardia

call the resulting data frame LaGuardiaFlights

2. LaGuardiaFlights should include the following variables

tenMinDelay: response {TRUE / FALSE}; indicates departure delay greater than 10 minutes;

remove cases with missing response

hour: hour of departure using 24 hr clock

weekend: {TRUE / FALSE} indicates flight departed on a Saturday or Sunday

AWND: average wind speed reported from the weather station at LaGuardia Airport

PRCP: precipitation reported from the weather station at LaGuardia Airport

SNOW: snowfall reported from the weather station at LaGuardia Airport

2

fog: {TRUE / FALSE} “Yes” if the weather station at LaGuardia Airport reported any of the

following:

– WT22: “Ice fog or freezing fog”

– WT01: “Fog, ice fog, or freezing fog (may include heavy fog)”, or

– WT02: “Heavy fog or heaving freezing fog (not always distinguished from fog”

Task 2.1.1 How many total cases are in the Training set?

Task 2.1.2 What percentage of these flights departed at least 10 minutes late?

Task 2.1.3 What percentage of these flights departed on the “Weekend”?

Task 2.1.4 What percentage of these flights departed on a day with “Fog”?

2.2 Partition test and training data

Task 2.2.1 partition 20% of the LaGuardiaFlights data for Test data and use the remaining 80%

for Training data (Note: make sure set.seed(822) is properly specified in the Front Matter of

your Rmd document)

2.3 Predictive modeling

Task 2.3.1 Fit a null model to the Training data, and report the accuracy of the model.

Task 2.3.2 Fit a logistic regression model to the Training data. Show the resulting confusion

matrix, and report the accuracy of the model.

Task 2.3.3 Fit a k-nearest neighbors model to the Training data. Show the resulting confusion

matrix, and report the accuracy of the model.

Task 2.3.4 Fit a naive Bayes model to the Training data. Show the resulting confusion matrix,

and report the accuracy of the model.

2.4 Visualization

** Task 2.4.1 Create a visualization that shows the predicted probability of 10 minute delay based on hour of

departure and precipitation using a logistic regression classifier. Specifically,**

The predicted probability of 10 minute delay should be clearly displayed for all plausible combinations

of hour and precipitation.

overlay Test data with clear indication of the observed response

demonstrate good plotting practices