CS373 Homework 1

Due date: Monday, February 11, 11:59pm (submit pdf on Gradescope)

Any use of late days must be explicitly mentioned at the top of your submission.

Homework must be submitted as a PDF; answers should be typed.

Instructions for submission
Submit a single PDF on Gradescope with all your answers. Make sure you
select the page corresponding to the beginning of each answer, else points
might be deducted. For part I, show the steps you took. For part II, include the R code
you used for analysis, along with its output and any plots required by the question. Please
label all plots with the question number. Your homework must be typed and must contain
your name and Purdue ID.

1 Part I: Basic Probability and Statistics

A. (6 pts) Consider an experiment where a fair die is rolled repeatedly until the first time
a 3 is observed.

(a) What is the sample space for this experiment? What is the probability that the
die turns up a 3 after i rolls?
(b) What is the expected number of times we roll the die?
(c) Let E be the event that the first time a 3 turns up is after an even number of
rolls. What set of outcomes belong to this event? What is the probability that E
occurs?

B. (5 pts) Two standard dice are rolled. Let E be the event that at least one of the dice
lands on 5; let F be the event that the sum of the dice is even; and let G be the event
that the sum is 7. Compute the following:

(a) P(E ∩ F )
(b) P(E ∪ F )
(c) P(E ∪ G)
(d) P(E ∩ ¬G)
(e) P(E ∪ F ∪ G)

C. (4 pts) We are given four coins C1 , C2 , C3 and C4 . Coin Ci is chosen randomly with
probability proportional to i for i = 1, 2, 3, 4. Let Hi represent the event that heads is
observed when Ci is tossed; P (H1 ) = 1/4; P (H2 ) = 1/3; P (H3 ) = 1/2; P (H4 ) = 2/3.

(a) What is the probability of selecting coin C3 ?

(b) If a coin is selected at random and tossed, find the conditional probability that
the coin is C3 given that a tail is observed. (State this as a conditional probability
and show the calculation.)

1
D. (4 pts) 52% of the students at a particular college are female. 5% of the students in the
college are majoring in computer science. 0.55% of the students are women majoring
in computer science.
(a) If a student is selected at random, find the conditional probability that the student
is female given that they are majoring in computer science. (State this as a
conditional probability and show the calculation.)
(b) If a student is selected at random, find the conditional probability that the student
is majoring in computer science given that they are female. (State this as a
conditional probability and show the calculation.)
(c) Now suppose that the overall proportion of female students increases to 57% and
that the conditional probability from D(a) changes (i.e., increases or decreases) to
15%. Compute the updated conditional probability that a student is majoring in
computer science given that they are female. (Assume that the overall proportion
of students majoring in CS stays the same.)
E. (6 pts) A system is built using 3 disks d1 , d2 , d3 having probabilities of failure 0.01,
0.03 and 0.05 respectively. Suppose the disks fail independently.
(a) Let W denote the event that the system will work, which happens if at least one
of the disks works. Compute P (W ), the probability that the system will work.
(b) Let A denote the event that at least one of the following happens: (i) d1 and d2
work; (ii) d3 works. If the system works when event A occurs, then compute the
probability that the system will work.
(c) Considering the setting of Eb, given that d1 works, what is the conditional prob-
ability that event A will occur and the system works?
F. (6 pts) Let an experiment consist of rolling three standard 6-sided dice.
(a) Compute the expected value of the sum of the rolls.
(b) Compute the variance of the sum of the rolls.
(c) If X represents the maximum value that appears in the three rolls, what is the
expected value of X?

2 Part II: R
In this assignment, you will use the R statistical package to explore, transform, and analyze
data. Based on your analysis you will formulate hypotheses about the data. To get started,
do the following:
• Download and install R from: http://cran.r-project.org/
• Download the Yelp dataset from Piazza.
This data set is part of the Yelp academic dataset and consists of data about 24,813
restaurants. The datafile yelp.csv contains 28 attributes: 6 numeric and 22 discrete.
The first row of the data file is a header row with the names of the attributes where
names are separated by a comma (,).

2
Use R to analyze the Yelp data and complete the questions below.

3 Data import and summarization

Read the data into R using read.table() function. Use the argument sep="," to specify
the column delimiter, the argument header=TRUE to read in the column names, the ar-
gument quote="\"" to read in the quoted fields, and the argument comment.char="" to
treat the # characters as text rather than comments.

(a) (2 pts) Print a summary of the data using the summary() function.

(b) (2 pts) Print the names of the columns in the table using the names() function.

4 1D plots
A. (a) (3 pts) Plot a histogram of the ‘reviewCount’ attribute. Use the hist() function
with its default values and make sure to title the plot with the name of the attribute
for clarity.
(b) (3 pts) Compute the logged values for ‘reviewCount’ (you can use log(d$column name)
to compute the log of all the values in a column). Plot a histogram of the logged
values.
(c) (3 pts) Plot a density plot of the logged values of the ‘reviewCount’ attribute using
the density() function.
(d) (3 pts) Discuss the similarities and differences between the three plots and the
information they convey about the distribution of ‘reviewCount’ values in the data.

B. (2 pts) Plot a barplot of the ‘state’ attribute to show the frequency of each value. Use
the table() function to get the counts for each value and the names() function to get
the names of the values in the table. Use the barplot() function with the names.arg
argument to label the bars with the appropriate value. Again, make sure to title the
plot with the name of the attribute for clarity.
(Note that this will look like a histogram but for nominal values. In small renderings
of this plot, you might not see all the state name labels, but if you stretch the window
you will be able to see all the labels.)

5 Sampling and transforming data

A. (4 pts) The attributes ‘categories’ and ‘recommendedFor’ each contain a comma sepa-
rated list of values associated with each restaurant. Compute two new boolean features:
‘servesPizza’ and ‘goodForBreakfast’ with a value of TRUE if the list contains Pizza
(in ‘categories’ ), breakfast (in ‘recommendedFor’ ) respectively and FALSE otherwise.
You can use the function grepl(str, f$column name) to check whether the values in
column name contain the string str.

3
 Append the two new columns to the original data frame, using cbind() to increase the
number of features by 2.

B. (a) (3 pts) Compute the quantiles (using quantile()) for the ‘checkins’ attribute.
(b) (3 pts) Select a subset of the data with ‘checkins’ value ≤ the 1st quartile (25th
percentile). You can use subset() or select from the data frame with [ ] opera-
tions.
(c) (3 pts) Print a summary of the above subset for the following attributes only:
‘checkins’, ‘stars’, ‘noiseLevel, ‘priceRange’, ‘reviewCount”, ‘goodForGroups’, and
compare them to their summary for the full dataset.

Discuss any differences that you find in the distributions of these attributes.

6 2D plots and correlations

A. (6 pts) Plot a scatterplot matrix (using plot()) for the five attributes:
‘stars’, ‘reviewCount’, ‘checkins’, ‘longitude’, ‘latitude’.

Identify which pair of attributes exhibit the most association (as you can determine
visually) and discuss if this is interesting or expected, given your domain knowledge.

B. (6 pts) Calculate the pairwise correlation among the above five attributes using the
cor() function.

Identify the pair of attributes with largest positive correlation and the pair with largest
negative correlation. Report the correlations and discuss how it matches with your
visual assessment in part A.

C. (6 pts) Plot a boxplot (using boxplot()) for each of the following four attributes vs.
the ‘attire’ attribute: ‘checkins’, ‘reviewCount’, ‘stars’, ‘latitude’.
Make sure to label both axes of the plot with the appropriate attribute names.

(a) Identify the attribute that exhibits the most association with ‘attire’ (as you can
determine visually) and discuss whether this is interesting or expected, given your
domain knowledge.
(b) For the attribute identified above, calculate its interquartile range for each value
of ‘attire’ (i.e., a separate IQR for the “casual” instances, the “dressy” instances,
the “formal” instances and the instances with “ ” for ‘attire’ ). You can do this
with the subset() and quantile() functions. Calculate the overlap between the
four IQRs. Discuss whether these results support the conclusion you made based
on visual inspection.

4
7 Identifying potential hypotheses (20 pts)
During your exploration above, investigate other aspects of the data. Explore relationships
between variables by assessing plots, computing correlation, or other numerical analysis.

Identify TWO possible relationships in the data (other than the ones specified in earlier
questions) and formulate hypotheses based on the observed data. For each of the two
identified relationships:

(a) Include a plot illustrating the observed relationship (between at least two variables).

(b) State whether the variables are discrete or continuous and what type of plot is relevant
for comparing these two types of variables.

(c) Formulate a hypothesis about the observed relationship as a function of two random
variables (e.g., X is associated with Y ).

(d) Write the hypothesis as a claim in English, relating it to the attributes in the data.

(e) Identify the type of hypothesis.