Khan Academy Ap CompSci Principles Data analysis

Ace your homework & exams now with Quizwiz!

Hong Lien is a research director for NASA. She is hoping to build a system that can process millions of images from satellites orbiting the earth and analyze them each day for signs of deforestation and ocean debris. She decides to hire an engineer specifically for the task of building the system and reviews many resumes. Which of these lines on a resume is most pertinent to this task?

"I have experience building highly scalable systems."

Part 1: In a 2019 research paper, a group of researchers analyzed two machine learning algorithms for automated hate speech detection. Both algorithms were trained on thousands of tweets that had been annotated by crowdsourced workers as being offensive or not. To test the algorithms, the researchers collected millions of tweets that were classified as written in either the dialect of African American English (AAE) dialect or White-aligned English. [How?] The researchers ran the algorithms over each tweet and recorded the rate at which non-offensive tweets were marked as offensive as well as the rate at which offensive tweets were marked as non-offensive. Algorithm 1AAEWhite-alignedLabeled offensive, but wasn't actually46.3%9.0%Labeled not offensive, but actually was1.1%7.9% Algorithm 2AAEWhite-alignedLabeled offensive, but wasn't actually26%4.5%Labeled not offensive, but actually was4.2%30.5% Which of the following statements is supported by that data?

1. Both algorithms contain bias that results in more often falsely labeling AAE tweets as "offensive" versus White tweets. 2.A discussion platform could calculate the number of posts that have been categorized as "hate speech" by the algorithm compared to the number of posts that were flagged by users.

Lakeisha is developing a program to process data from smart sensors installed in factories. The thousands of sensors produce millions of data points each day. When she ran her program on her computer, it took 10 hours to complete. Which of these strategies are most likely to speed up her data processing? 👁️Note that there are 2 answers to this question.

1. Distributing the computing to multiple machines to run the program on subsets of the data. 2.Using parallel computing on a computer with a multi-core CPU.

Oil tankers are prone to fire, due to all the gasoline on the ship. Some tankers use specialized cameras for flame and smoke detection, and install them in the most fire-prone spots of the ship. The cameras record videos whenever they detect motion, and also record metadata along with the videos. The metadata includes the location of the camera, the temperature near the camera, the start date/time of the recording, and the end date/time of the recording. Which of these questions can be better answered by analyzing the metadata instead of the recorded videos? 👁️Note that there are 2 answers to this question.

1. On average, how many recordings are made each day per camera location? 2. What is the range in temperature at the cameras?

Two neighboring cities have created data sets of places where people can get their flu shot. Mountain View stores the data in this format: Columns:Facility name, Street address, Zip code, Start date, End dateSample row:Southeast Health Clinic, 2420 Shotwell St, 94041, 2013-11-22, 2013-11-25 Los Altos stores the data in this format: Columns:Facility name, Street address, Begin date, End date, EligibilitySample row:Public Health Center, 1301 Pierce St, 2013-11-15, 2013-12-05, Uninsured adults The two cities are combining their data sets to create informational campaigns for their residents. Which of the following can be determined from the combined data sets? 👁️Note that there are 2 answers to this question.

1. The last possible date get a flu shot 2. The city that has the most locations opne

Part 1: In 2016, researchers studied the differences between two dialects of English used in tweets on Twitter. They first categorized over eight million tweets as either African-American-aligned (AA-aligned) or White-aligned. [How?] The researchers then tried out three different language identification algorithms to see whether each tweet would be categorized as the English language, a different language, or an unknown language. The following table shows the proportion of tweets from each dialect that were classified as non-English by each algorithm: AA-alignedWhite-alignedAlgorithm 113.2%7.6%Algorithm 28.4%5.9%Algorithm 324.4%17.6% Which of the following statements is supported by that data?

1.All three algorithms contain biases that cause them to misidentify an AA-aligned tweet as non-English more often than they misidentify the language of a White-aligned tweet. 2.The analytics team of a reviews website uses the language identification algorithm to display a pie chart of the number of reviews written per language in the last year.

The San Francisco Health Department keeps track of health inspections at restaurants and makes the data publicly available. Each row in the inspections data set contains these details: Restaurant name Restaurant address Inspection date Inspection score (0-100) Violation description Risk severity (low/medium/high) Which of the following questions can be answered using the available data? 👁️Note that there may be multiple answers to this question.

1.How many restaurants have an inspection score greater than 90? 2.What is the average inspection score for the high risk violations? 3.Which restaurant has the lowest inspection score?

Community gardens are public gardens where local residents can grow plants in a plot. They are very popular, so there are often waitlists to get a plot. Alioto Community Garden stores their waitlist data in this format: Columns:Name, email, address, waitlist date, plot sizeSample row:Jolie Clover, [email protected], 501 Stanyan St, 05-06-2018, small A neighboring garden, Arkansas Friendship Garden, stores their waitlist data in this format: Columns:Last name, first name, phone, address, waitlist dateSample row:McGee, Eirene, 631-421-4141, 1351 24th Ave, 11-11-2018 The gardens decide to combine their data sets, since they're located so near to each other. Which of the following can be done using the combined data set?

1.Make a map of the waitlisted people 2.Figure out who has been waiting the longest

Note that this is a two-part question. Part 1: Yitaf discovers a new startup called GenderChecker, a service that claims to be able to "identify the gender of your customers" based on an email address. Yitaf is interested in the accuracy of its predictions and tries out a bunch of addresses: InputProbability of maleProbability of [email protected]%57.2%[email protected]%10.3%[email protected]%57.8%[email protected]%1.1%[email protected]%54.1%[email protected]%17.8% Which of the following statements is most supported by that data? Choose 1 answer:Choose 1 answer:

1.The algorithm contains bias that associates "doctor" (and related abbreviations) with the male gender. 2.A social media network could generate a graph that shows the fraction of user sign-ups each month per gender.

Bianca is planning to start a service for programmers who want to prepare for software engineering interviews. To help her figure out the target audience, she does some market research by sending around a survey. The survey asks: How many years have you been programming? From 1-10, how interested are you in a service that helps you prepare for interviews? How much would you be willing to pay monthly for the service? She creates two scatter plots based on the results. The first plot compares years of programming experience to interest in the service: \small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{interest}interest\text{years}years A scatter plot with years on the x axis and interest on the y axis. Both axes go from 0 to 10. Dots are scattered on the plot, starting in the upper left corner (e.g. [1, 9]) and generally sloping down towards the bottom right corner (e.g. [7.5, 2]). The second plot compares willing payment amount to interest: \small{10}10\small{20}20\small{30}30\small{40}40\small{50}50\small{60}60\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{interest}interest\text{payment}payment A scatter plot with payment on the x axis and interest on the y axis. The x axis goes from 0 to 70 and the y axis goes from 0 to 10. Dots are scattered on the plot, starting in the bottom left corner (e.g. [12, 2]) and generally sloping up towards the top right corner (e.g. [60, 9]). Which conclusions can Bianca make from the data? 👁️Note that there are 2 answers to this question.

1.There is a negative correlation between years of programming experience and interest in the service. 2.A higher interest in the service is positively correlated with a higher willingness to pay.

A "red light camera" is a camera installed at street intersections that records whenever a car runs a red light. The camera records two images, one right before the car enters the intersection, and one after it's entered the intersection. In addition to the images, it records metadata about the incident: the date and time, the intersection location, the speed of the car, and the seconds elapsed past the light turning red. Which of these questions can be better answered by analyzing the metadata instead of the image data? 👁️Note that there are 2 answers to this question.

1.What is the average speed of a car when it runs a red light? 2.Which intersections have the greatest number of red light runners?

An obstetrics department is studying fetal heartbeat and how it corresponds to a healthy birth. They make audio recordings of the fetal heartbeat at various stages of pregnancy. Along with each recording, they also record metadata. The metadata includes the gestational age of the fetus (in weeks), the age of the mother, the height of the mother and the weight of the mother. Which of these questions can be better answered by analyzing the audio data instead of the metadata?

1.What is the range in the heartbeat of a fetus? 2.What is the average heartbeat of a fetus?

The Chicago Police Department uses a database to keep track of reported crimes. After anonymizing the data, they make it freely available online. Here's what the the crime data set includes: The date of the crime The address (block level) The type of crime (theft/battery/robbery/assault/etc.) The location type (street/residence/business/etc.) Whether an arrest was made (true/false) Which of the following questions can be answered using the available data? 👁️Note that there may be multiple answers to this question.

1.What type of crime was the most common for each year in the data set? 2.Which month has the most number of robberies? 3.What was the average number of crimes committed per location type?

Tajikstan is a country in Central Asia, where many people live in poverty and most do not have access to the Internet. The following table shows the percentage of residents using the Internet for the years 2012-2017: Year% Internet usage201214.51\%14.51%14, point, 51, percent201316.00\%16.00%16, point, 00, percent201417.49\%17.49%17, point, 49, percent201518.98\%18.98%18, point, 98, percent201620.47\%20.47%20, point, 47, percent201721.96\%21.96%21, point, 96, percent Source: International Telecommunications Union Assuming the Internet usage keeps growing at a similar rate, which of the following is the most reasonable prediction for Internet usage in 2019 (two years after the last data point)?

25.1%

Greenland is the world's largest island, located east of Canada. It's connected to the rest of the world's Internet via underwater fiber cables. The following table shows the percentage of Greenland residents using the Internet for the years 2012-2017: Year% Internet usage201264.90%64.9064, point, 90201365.80%65.8065, point, 80201466.70%66.7066, point, 70201567.60%67.6067, point, 60201668.50%68.5068, point, 50201769.48%69.4869, point, 48 Source: International Telecommunications Union Assuming the Internet usage keeps growing at a similar rate, what is the most reasonable prediction for Internet usage in 2019 (two years after the last data point)?

71.4%

The Democratic Republic of Congo is a country located in Central Africa, where many people live in extreme poverty and few have access to the Internet. The following table shows the percentage of residents using the Internet for the years 2013-2017: Year% Internet usage20136.60\%6.60%6, point, 60, percent20147.11\%7.11%7, point, 11, percent20157.62\%7.62%7, point, 62, percent20168.12\%8.12%8, point, 12, percent20178.65\%8.65%8, point, 65, percent Source: International Telecommunications Union Assuming the Internet usage keeps growing at a similar rate, what is the most reasonable prediction for Internet usage in 2019 (two years after the last data point)?

9.8%

Company recruiters use applicant tracking systems to keep track of the résumés that candidates send in for the job. Many applicant tracking systems use algorithms to automatically rank the résumés, to help recruiters sift through large quantities of résumés. Which of these algorithms would require access to a large database of résumés?

An algorithm that ranks résumés based on similarity to résumés from already hired applicants.

Talisa is an engineer that is helping a museum to digitize and analyze all of its historical books. After running the software over the first 100 books, she realizes that the museum computer has run out of space to store the digital files. Which technique is the most needed to help them digitize the remaining books?

Distributed computing

A national bank opts to use machine learning for deciding whether to award loans to applicants. The engineers create the algorithm by training a neural network on their large database of previous loan applications and decisions (made by loan officers). After they start using the algorithm for new loan applicants, they receive complaints that their algorithm must be biased, because all the loan applicants from a particular zip code are always denied. What is the most likely explanation for the algorithm's bias?

For that zip code, the training data set only has loan applications that were denied.

Andy is using machine learning for an algorithm that classifies photos of restaurant meals by category (such as "sandwich", "curry", or "salad"). He trains a neural network on a large open database of photos of restaurant meals. He then tests the network on local restaurants and notices that the Ethiopian restaurant meals aren't classified correctly. What's the best way to improve the machine learning algorithm's ability to recognize Ethiopian meals?

He can add Ethiopian meals to the training data set, by finding photos online, crowd-sourcing, or taking them himself.

A hospital IT department is determining how much data storage capacity they will need to store electronic health records for patients. They start by making a list of the type of data that comes from each department: DepartmentDataFormat & average sizePrimary careNotes from patient chats with doctors3 paragraphs per visitLaboratoryTest resultsA table with 20 rows and 3 columnsRadiologyImagery from scans (CT/PET/MRI)64 1024x1024 grayscale imagesPharmacyMedication prescriptionsPatient name/ID, medicine name, data Which type of data is likely to require the most data storage capacity?

Imagery from scans (CT/PET/MRI)

An online curriculum provider offers their product to two audiences: independent learners (self-directed) and classroom learners (led by their teachers). They want to understand the differences between the audiences and how they use the product, so they sent surveys and collected data. Users rated their satisfaction with the product from 1-10, where 1 is least satisfied and 10 is the most satisfied. This scatter plot compares the hours per week spent by a user to their rating of the product: \small{1}1\small{2}2\small{3}3\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{rating}rating\text{hours/week}hours/week\purpleE{\text{independent}}independent\greenE{\text{classroom}}classroom A scatter plot with hours per week on the x axis and rating on the y axis. The x axis goes from 0 to 4 and the y axis goes from 0 to 10. Both purple and green dots are scattered on the plot. The purple dots start in the bottom left corner (e.g. [0.2, 2]) and generally slope up towards the top right corner (e.g. [3.5, 9]). The green dots are scattered on the plot for all the the x axis values between the rating values of 4 and 7, seemingly randomly. The \greenE{\text{green dots}}green dotsstart color #0d923f, start text, g, r, e, e, n, space, d, o, t, s, end text, end color #0d923f represent classroom learners and the \purpleE{\text{purple dots}}purple dotsstart color #543b78, start text, p, u, r, p, l, e, space, d, o, t, s, end text, end color #543b78 represent independent learners.

Independent learners are generally more satisfied with the product as their usage increases.

A travel website is adding a feature for users to store trip itineraries. Here's a sample itinerary: Title: Summer trip to Japan 1. Inari shrine (Kyoto, Japan) 2. Iwatayama Monkey Park (Kyoto, Japan) 3. Fushimi Inari Taisha (Kyoto, Japan) 3. Fukui Prefectural Dinosaur Museum (Katsuyama, Japan) 4. Kōtoku-in (Kamakura, Japan) 5. Ghibli Museum (Mitaka, Japan) 6. Tokyo Anime Center (Tokyo, Japan) They are considering a number of enhancements to the trip itinerary feature, and the engineering team is considering the data storage requirements of the new features. Which feature is likely to require the greatest increase in data storage needs?

Making copies of the user's trip itinerary in 6 data centers around the world

Xiomara is a researcher studying the effect of carbon emissions from airplanes on global warming. She collects millions of data points tracking the path of airplanes and develops a program that analyzes the data. When she runs the program on a single company's airplanes, it takes an hour to complete, so she becomes concerned that it will take much too long to run on all of the airplane data. Her friend Dacari suggests using parallel computing to speed up the analysis of the airplane emission data. How would parallel computing speed up the analysis?

Parallel computing can run the program in parallel on subsets of the data, so that the total amount of time is less.

A team of scientists and engineers is putting together a research project to study whale sounds. In order to develop the infrastructure for the project, they need to first determine how much data storage space their observational data will require. This is an example of a single observation: SoundLocationDate/timeSpecies3 minute long audio file63.776871, -171.742193May 27, 2019, 2:23:13 PMBeluga The team hopes to capture thousands of whale sounds from all the world's oceans. Which piece of data will increase their data storage needs the most?

Recording of whale sound

Brittany is using machine learning for an algorithm that classifies social media posts according to their sentiment ("positive", "negative", or "neutral"). She trains a neural network on a large open database of social media posts and tests the network on her personal social media feed. She notices that it's mis-classifying the posts from her teenage friends, who use different slang from her other friends. What's the best way that Brittany can improve the machine learning algorithm's ability to classify posts from teenagers?

She can add social media posts from teenagers into the training data set, both from her own network and globally available data.

Safiya is a software engineer at a company that's developing software for self-driving cars. She's working on software that uses computer vision and machine learning algorithms to detect pedestrians walking near the car and trigger the brakes when needed. After training the algorithm on a large dataset of training data (videos of pedestrians walking near cars), they try it out in cars with backup drivers. The drivers report that it detected most pedestrians, but failed to detect people using wheelchairs and parents pushing strollers. The drivers had to manually respond in those cases. What's the best way that Safiya can improve the machine learning algorithm's ability to detect all pedestrians?

She can add videos of people using wheelchairs and strollers into the training data set (perhaps crowd-sourcing them if there are none already available).

On June 22, 1944, the U.S. introduced the G.I. Bill, a law that provided many benefits to war veterans, including college tuition. Cornell University has been tracking enrollment numbers since their inception. This table shows enrollment in the 10-year period from 1940-1950, broken down by gender: YearMale enrollmentFemale enrollmentTotal enrollment19405,5701,5467,11619415,2991,6476,94619424,7891,6906,47919433,1281,7484,87619442,7222,1124,83419453,1412,2025,34319467,3581,8919,24919477,8641,9379,80119487,9011,8529,75319497,9491,8959,84419507,8571,9719,828 Source: Cornell Institutional Research & Planning Which hypothesis is most consistent with the data?

The G.I. Bill led to a large increase in male enrollment.

HireView is a startup that claims to speed up the process of interviewing job candidates. A candidate submits a video answering interview questions and HireView analyzes the video with a machine learning algorithm. The algorithm scores the candidate on various aspects of their personality, such as "willingness to learn" and "personal stability". HireView engineers trained the algorithm on a data set of past videos that were scored by employers and psychologists. In a test of the algorithm, HireView engineers discover that the algorithm always gives lower scores to people who speak more slowly. What is the most likely explanation for the algorithm's bias?

The algorithm was trained on data where videos with slower speech were scored lower, due to the bias of the scorers.

A medical diagnosis app lets users track their symptoms. Whenever a user reports a symptom, the app adds a row to a database table. Each row contains: The user ID The date of the report The time of the report A description of how they're feeling The severity of the feeling (1-10) Here are a few rows from the table: user_iddatetimedescriptionseverity6203811/19/201807:52Skin rash on arms42039409/24/201803:45Pounding headache93691704/11/201823:22Leg cramps2 The app marketing team wants to understand their users better and asks the data analyst for various statistics. Which statistic can not be calculated from the table of reports?

The average duration of the feeling.

A non-profit website decided to launch a fundraising campaign in December, to encourage people to make tax-deductible donations before the end of the year. For their campaign, the website displayed a "Please donate!" banner along the top of every page, starting on December 15th. This table shows their data from December 12th to December 22nd, tracking donations from signed in users, donations from users that weren't signed in, and total sign-ups for the site. DayDonations (Signed in)Donations (Signed out)Sign-ups12/12$11,775$2,02483112/13$11,783$2,52787412/14$11,455$2,84983912/15$22,582$3,73286412/16$22,867$3,72485312/17$22,669$3,81089312/18$23,679$3,27089712/19$23,577$3,47780312/20$23,866$3,05284212/21$23,837$3,63481112/22$24,519$3,928855 Which hypothesis is most consistent with the data? Choose 1 answer:Choose 1 answer:

The donation banner led to a significant increase in total donations and did not affect sign ups.

An online toy store keeps a database of all sales. For each purchase, the database includes the following details: The date of the sale The time of the sale The method of payment (credit/PayPal) The total amount paid A list of the items sold Here are a few rows from the database: datetimemethodtotalitems01/09/201914:44PayPal38.47"Play-doh 36-pack", "MagicBeadz"02/13/201911:25credit18.19"Grumpy cat stickers", "Sidewalk chalk"03/04/201918:13PayPal59.42"Lego Death Star", "Etch-a-sketch" The toy store manager asks the database administrator for a report on various sale metrics. Which of these metrics can not be reported from the sales database?

The most expensive item sold.

A website for sports fans includes a discussion forum for fans to discuss games, athletes, and predictions for the coming seasons. They decide to redesign their discussion forum to be more usable and modern looking, and ask the data analysis team to analyze the effect of the redesign on usage statistics. The website released the redesign on March 8, 2018. This table shows daily usage data before and after the redesign, including number of posts created, number of replies, and number of upvotes on posts: DayPostsRepliesUpvotes3/1/18592132840643/2/18560134941913/3/18555136240843/4/18576131841213/5/18582133240033/6/18559134040303/7/18576132340663/8/18591155831033/9/18552159930443/10/18587161130733/11/18581156730893/12/18557155430153/13/18565160730503/14/1859916013062 Which hypothesis is most consistent with the data?

The redesign led to a significant increase in replies and a decrease in upvotes.

Shameeka is setting up a computing system for predicting earthquakes based on processing data from seismographs (devices that record earth movements). The system will start off with data from local seismographs but eventually handle millions of data points from seismographs worldwide. For her system to work well, what is an important feature?

The system must be scalable.

StackOverflow is a popular question & answers site. Each time a user asks a new question, they insert a row in a database table. Each row contains: The user ID The user display name The timestamp of the question The text of the question The spam score of the question (0-5) Here are a few rows from the table: user_iddisplay_nametimestampquestionspam_score62038TheAskerator11/27/2012 06:15:28How do I geocode a lat/lng?020394NewCoder12303/12/2015 10:55:10Where can I host my website for free?136917QuestionErrthing05/04/2014 11:34:25Wanna download this free file?3 The team wants to display question statistics on an internal dashboard. Which statistic can not be calculated from the table of questions?

The user ID with the most number of unanswered questions

Craig is developing a new micro-blogging app and has shared it with a group of beta testers. He wants to understand their usage patterns, so he tracks data on the number of posts a user makes each day, the average length of their posts, and the average sentiment of their posts (from very negative to very positive). This plot compares the average sentiment for a user's posts to the number of posts they make each day: \small{0.5}0.5\small{1}1\small{\llap{-}0.5}-0.5\small{\llap{-}1}-1\small{5}5\small{10}10\small{15}15\small{20}20\text{posts/day}posts/day\text{sentiment}sentiment A scatter plot with sentiment on the x axis and posts per day on the y axis. The x axis goes from -1 to 1, and the y axis goes from 0 to 20. Dots are scattered on the plot, starting in the upper left corner (e.g. [-1, 20]) and generally sloping down towards the bottom right corner (e.g. [0.9, 1]). This second plot compares the average length of a user's posts to their number of posts per day: \small{40}40\small{80}80\small{120}120\small{160}160\small{200}200\small{240}240\small{5}5\small{10}10\small{15}15\small{20}20\text{posts/day}posts/day\text{post length}post length A scatter plot with post length on the x axis and posts per day on the y axis. The x axis goes from 0 to 280, and the y axis goes from 0 to 20. Dots are scattered on the plot for all values of post length and posts per day, seemingly randomly. Which conclusions can Craig make from the data?

There is a negative correlation between posts per day and sentiment.

A mood tracking app decides to help users understand their mood changes better by also tracking the hours they spend on other applications. This chart visualizes the results for a video watching app, using a scatter plot to compare each user's hours spent in the app to their mood after exiting the app: \small{1}1\small{2}2\small{3}3\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{mood}mood\text{hours}hours\purpleE{\text{amusement}}amusement\greenE{\text{education}}education A scatter plot with hours on the x axis and mood on the y axis. The x axis goes from 0 to 4 and the y axis goes from 0 to 10. Both purple and green dots are scattered on the plot. The purple dots start in the upper left corner (e.g. [0.2, 9]) and generally slope down towards the bottom right corner (e.g. [3.5, 2]). The green dots are scattered on the plot for all the the x axis values between the mood values of 6 and 8, seemingly randomly. Users rate their mood from 1-10, where 1 is least happy and 10 is most happy. The \greenE{\text{green dots}}green dotsstart color #0d923f, start text, g, r, e, e, n, space, d, o, t, s, end text, end color #0d923f represent users that reported using the app primarily for educational videos, and the \purpleE{\text{purple dots}}purple dotsstart color #543b78, start text, p, u, r, p, l, e, space, d, o, t, s, end text, end color #543b78 represent users that reported using it primarily for amusing videos.

Users who watch amusing videos generally feel less happy the more they watch.


Related study sets

I-core marketing practice questions 1,5,2,6,9,11

View Set

ASNT Level 3 Basic Materials & Processes

View Set

ES - #4 - WATER POLUTION - CHAPTER 11

View Set

Top 5 Hardest Geography questions (for you Anton) NO CHEATING!!!!!!!!!!1

View Set

DMV Questions (NO SIGN PICTURES)

View Set

Modules 1 - 4: Switching Concepts, VLANs, and InterVLAN Routing Exam

View Set