Here are the questions and their short expansions:

Probability and Probability Distribution

1.​ What is the R function to compute the density of the normal distribution?
○​ The R function is dnorm(x, mean, sd). It calculates the probability density at a given
point x for a normal distribution with specified mean and standard deviation (sd).
2.​ How do you calculate the cumulative probability of a value under the normal curve?
○​ You use the pnorm(x) function. This function gives the probability that a random
variable from a standard normal distribution is less than or equal to x.
3.​ Which function returns quantiles from a normal distribution?
○​ The qnorm(p) function returns quantiles. It takes a probability p and returns the
value x such that the cumulative probability up to x is p.
4.​ Write an R command to compute the probability of getting 3 successes in 5 trials
with p=0.6.
○​ The R command is dbinom(3, size = 5, prob = 0.6). This uses the binomial
probability mass function to calculate the exact probability of 3 successes in 5 trials.
5.​ What does the rbinom() function do?
○​ The rbinom() function generates random binomial values.
6.​ How do you find the probability of a value greater than 1.5 in a standard normal
distribution?
○​ You use 1 - pnorm(1.5). Since pnorm() gives the cumulative probability up to a
value, subtracting it from 1 gives the probability of being greater than that value.
7.​ What does qnorm(0.95) return?
○​ It returns the 95th percentile of the standard normal distribution. This means 95% of
the values in a standard normal distribution fall below this value.
8.​ Use R to generate the first 5 probabilities of a Poisson distribution with lambda = 2.
○​ The R command is dpois(0:4, lambda = 2). This calculates the probability mass for
X = 0, 1, 2, 3, 4 for a Poisson distribution with a rate parameter of 2.
9.​ What is the purpose of dbinom() in R?
○​ dbinom() computes the binomial probability mass function. It gives the probability of
exactly x successes in n trials.
10.​Write the syntax to compute the tail probability (P(X>5)) of a Poisson(3) variable.
○​ The syntax is 1 - ppois(5, lambda = 3). Similar to the normal distribution, subtracting
the cumulative probability from 1 gives the upper tail probability.
Generating Random Numbers
1.​ Generate 20 random values from a uniform distribution between 5 and 15.
○​ The R command is runif(20, min = 5, max = 15).
2.​ How do you set a seed for reproducibility?
○​ You use set.seed(123). Setting a seed ensures that random number generation
produces the same sequence of numbers every time the code is run.
3.​ Write R code to generate 10 random values from an exponential distribution with
rate = 0.2.
○​ The R code is rexp(10, rate = 0.2).
4.​ What is the difference between runif() and rnorm()?
○​ runif() generates random values from a uniform distribution, while rnorm()
generates random values from a normal distribution.
5.​ How to simulate 100 values from a normal distribution with mean = 50, sd = 5?
○​ You use rnorm(100, mean = 50, sd = 5).
6.​ Generate 50 Bernoulli trials with probability 0.4.
 ○​ The R command is rbinom(50, size = 1, prob = 0.4). A Bernoulli trial is a binomial
distribution with size = 1.
7.​ What does sample() do in R?
○​ sample() randomly selects elements from a vector or sequence.
8.​ How can you draw 15 values from a normal distribution and ensure you get the
same result each time?
○​ You set a seed first: set.seed(42); rnorm(15).
9.​ How do you simulate random data from a geometric distribution?
○​ You use rgeom(n, prob).
10.​Generate 100 random integers between 1 and 20.
○​ The R command is sample(1:20, 100, replace = TRUE). replace = TRUE is used
because you are drawing multiple times and want to allow for the same number to
be drawn again.
Selecting Random Samples
1.​ How to sample 5 elements from a vector v without replacement?
○​ You use sample(v, 5, replace = FALSE). replace = FALSE ensures that once an
element is selected, it cannot be selected again.
2.​ What is the function to sample rows from a data frame?
○​ You can use sample_n(df, n) or df[sample(nrow(df), size), ].
3.​ Write code to select a 10% sample of a dataset df.
○​ You can use sample_frac(df, 0.1).
4.​ How do you do stratified sampling in R?
○​ You use group_by() and sample_n() from the dplyr package. This allows you to
sample a certain number of observations from each group.
5.​ What is the difference between sampling with and without replacement?
○​ Sampling with replacement allows duplicates (an element can be selected multiple
times), while sampling without replacement does not (once an element is selected,
it's removed from the pool).
6.​ Select a sample of 30 students randomly from a list of 100.
○​ The R command is sample(1:100, 30). By default, sample() samples without
replacement when a single integer x is provided as the first argument, and size is
less than or equal to x.
7.​ How can dplyr be used for random sampling?
○​ You use sample_n() from the dplyr package.
8.​ How to sample from multiple groups in a data frame?
○​ You use group_by(group_var) %>% sample_n(5). This will sample 5 observations
from each unique value in the group_var column.
9.​ What does replace=TRUE mean in the sample() function?
○​ replace=TRUE means that sampling is done with possible repetitions.
10.​Sample 3 rows at random from the mtcars dataset.
○​ The R command is mtcars[sample(nrow(mtcars), 3), ].
Empirical Study of the Sampling Distribution of the Estimators
1.​ Simulate the sampling distribution of the mean of 100 random normal values,
repeated 500 times.
○​ The R command is replicate(500, mean(rnorm(100))). This generates 100 random
normal values, calculates their mean, and repeats this process 500 times.
2.​ How can you use replicate() to run a simulation multiple times?
○​ replicate() repeats an expression multiple times.
 3.​ What is the empirical sampling distribution?
○​ It is a distribution of a statistic (like the mean or median) obtained from repeated
samples.
4.​ How to visualize the distribution of simulated means?
○​ You can use hist() or ggplot().
5.​ How do you compute the standard error from the simulated means?
○​ You compute sd(simulated_means). The standard deviation of the simulated means
is an estimate of the standard error of the mean.
6.​ Why is replication important in simulation?
○​ Replication is important to approximate distributional properties. By repeating the
simulation many times, you can observe the behavior of the statistic.
7.​ What estimator properties can be studied through simulation?
○​ Bias, variance, and consistency can be studied through simulation.
8.​ Write code to study the sampling distribution of the median.
○​ The R code is replicate(1000, median(rnorm(50))). This simulates the median of 50
normal values, repeated 1000 times.
9.​ How do you analyze the bias of an estimator using simulation?
○​ You compare the mean of the sampling distribution to the true value of the
parameter.
10.​Compare the sampling distributions of the mean and median for a skewed
distribution.
○​ You simulate both and compare them using hist() or boxplot().
Simulation of Data from a Probability Distribution
1.​ Simulate 200 values from a normal distribution with mean 0 and sd 1.
○​ The R command is rnorm(200). By default, rnorm() generates from a standard
normal distribution (mean=0, sd=1).
2.​ How to generate 100 values from a binomial distribution with n=10, p=0.5?
○​ You use rbinom(100, size = 10, prob = 0.5).
3.​ Simulate 50 values from an exponential distribution.
○​ You use rexp(50).
4.​ Generate a histogram of 1000 simulated values from a normal distribution.
○​ The R command is hist(rnorm(1000)).
5.​ What is the difference between dnorm and rnorm?
○​ dnorm is for density (probability density function), while rnorm is for random
generation (generating random numbers from the distribution).
6.​ Simulate data from a uniform distribution between 0 and 10.
○​ The R command is runif(100, 0, 10). (Assuming 100 values, though not specified in
the question, a quantity is needed for simulation).
7.​ How to simulate from a chi-square distribution in R?
○​ You use rchisq(n, df).
8.​ How to simulate values from a beta distribution?
○​ You use rbeta(n, shape1, shape2).
9.​ Simulate 30 values from a negative binomial distribution.
○​ You use rnbinom(30, size, prob).
10.​What is the purpose of simulating data from known distributions?
○​ The purpose is to model and test theoretical properties. It allows you to understand
how statistical methods behave under known conditions.
Simulation of Data for a Regression Model
 1.​ Simulate linear data with y=4+3x+error where error N(0,1).
○​ The R code is x <- rnorm(100); y <- 4 + 3 * x + rnorm(100). This creates x values
from a normal distribution and then y values based on the linear model with added
normal error.
2.​ How do you fit a linear regression model in R?
○​ You use lm(y ~ x).
3.​ What is the role of the error term in simulated regression data?
○​ The error term introduces variability, similar to real-world data.
4.​ Simulate a dataset with two independent variables.
○​ The R code is x1 <- rnorm(100); x2 <- rnorm(100); y <- 5 + 2 * x1 - x2 + rnorm(100).
5.​ What R function is used to create a model matrix?
○​ The model.matrix(~ x1 + x2) function is used.
6.​ Simulate data where y depends non-linearly on x.
○​ An example is y <- x^2 + rnorm(100). This simulates a quadratic relationship.
7.​ How do you add multicollinearity into your simulated regression data?
○​ You simulate x1 and x2 to be highly correlated. For example, x2 <- x1 + rnorm(100,
sd = 0.1).
8.​ Fit a multiple linear regression model on simulated data.
○​ You use lm(y ~ x1 + x2).
9.​ How to simulate heteroskedastic errors?
○​ You can use rnorm(n, mean = 0, sd = abs(x)). This makes the standard deviation of
the error term dependent on the value of x.
10.​Visualize residuals of the simulated model.
○​ You use plot(residuals(model)).
Simulation of Data for Time Series Model
1.​ Simulate a random walk time series.
○​ You use cumsum(rnorm(100)). A random walk is the cumulative sum of random
steps (e.g., normal random values).
2.​ How to simulate an AR(1) model with phi = 0.8?
○​ You use arima.sim(n = 100, list(ar = 0.8)).
3.​ What does the function arima.sim() do?
○​ arima.sim() simulates from ARIMA (AutoRegressive Integrated Moving Average)
models.
4.​ Simulate a MA(1) time series.
○​ You use arima.sim(n = 100, list(ma = 0.5)).
5.​ Plot an AR(2) time series with specified coefficients.
○​ You use arima.sim(n = 100, list(ar = c(0.5, 0.3))). (The question asks to plot, but the
answer provides the simulation code).
6.​ Simulate and plot a seasonal time series.
○​ You can use sine waves plus noise: y <- sin(1:100/5) + rnorm(100). (The question
asks to plot, but the answer provides the simulation code).
7.​ How to introduce trend and noise in time series simulation?
○​ You can add a linear trend component and noise: y <- 0.1 * (1:100) + rnorm(100).
8.​ Generate a white noise series.
○​ You use rnorm(100). White noise is a series of independent and identically
distributed random variables.
9.​ What does an ACF plot of a simulated AR(1) look like?
○​ It shows slowly decaying autocorrelations.
 10.​How to simulate a time series with both trend and seasonality?
○​ You combine trend and seasonal parts with noise. For example, y <- 0.1 * (1:100) +
sin(1:100/5) + rnorm(100).
Monte Carlo Simulation
1.​ Estimate pi using Monte Carlo simulation.
○​ The R code is:​
n <- 10000;​
x <- runif(n); y <- runif(n)​
4 * mean(x^2 + y^2 <= 1)​
This simulates random points within a unit square and checks if they fall within a
quarter circle.
2.​ What is the basic idea behind Monte Carlo simulation?
○​ It involves estimating quantities via repeated random sampling.
3.​ Simulate the probability of getting at least one six in 4 dice rolls.
○​ The R command is mean(replicate(10000, any(sample(1:6, 4, replace = TRUE) ==
6))). This simulates 4 dice rolls many times and checks if any roll is a 6.
4.​ Use Monte Carlo to estimate the area under a curve.
○​ You simulate x values and compute the mean of the function values.
5.​ What R function is helpful for repeating simulations?
○​ The replicate() function is helpful.
6.​ How does increasing the number of repetitions affect the accuracy?
○​ Accuracy improves with more repetitions.
7.​ Simulate the mean of uniform random variables and analyze convergence.
○​ The R code is means <- replicate(10000, mean(runif(100))) and then plot(means) to
visualize convergence.
8.​ Estimate probability of winning a game using Monte Carlo.
○​ You model the game logic and repeat it many times.
9.​ What is variance reduction in Monte Carlo methods?
○​ These are techniques like control variates and antithetic variates that aim to reduce
the variance of the Monte Carlo estimate without increasing the number of
simulations.
10.​Use Monte Carlo to evaluate an integral.
○​ You can estimate a definite integral via uniform sampling.
Bootstrapping
1.​ What is bootstrapping used for in statistics?
○​ Bootstrapping is used to estimate the sampling distribution of a statistic using
resamples from the observed data.
2.​ How to perform bootstrap sampling of a vector in R?
○​ You use sample(x, replace = TRUE).
3.​ Estimate the standard error of the mean using bootstrapping.
○​ The R code is:​
means <- replicate(1000, mean(sample(x, replace = TRUE)))​
se <- sd(means)​
This generates many bootstrap samples, calculates the mean for each, and then
takes the standard deviation of these bootstrap means.
4.​ Plot a histogram of the bootstrap means.
○​ You use hist(means).
5.​ Calculate the 95% CI using the percentile method.
 ○​ You use quantile(means, c(0.025, 0.975)). This finds the 2.5th and 97.5th
percentiles of the bootstrap means.
6.​ What is the role of replacement in bootstrapping?
○​ Replacement allows drawing with replacement, mimicking resampling from the
original population.
7.​ How does bootstrapping differ from Monte Carlo simulation?
○​ Monte Carlo simulates theoretical models (often from known distributions), while
bootstrapping resamples from the observed data itself.
8.​ Use boot() from the boot package to bootstrap the median.
○​ You use boot(data = x, statistic = function(x, i) median(x[i]), R = 1000).
9.​ How many bootstrap samples are usually recommended?
○​ Typically 1000-10000 samples are recommended.
10.​What are the limitations of bootstrapping?
○​ It doesn't always work well with small samples or highly skewed data.