Scribd
Upload
8 views

R Problem Solving

This document discusses modeling Markov chains in R. It provides examples of creating a transition matrix in R and using packages like "diagram" and "Markovchain" to plot the state space and calculate properties of the Markov chain. It then shows an example of using a transition matrix to find the expected number of coin flips needed to get three heads in a row, modeling it as a Markov chain with states for the number of consecutive heads seen so far.

Uploaded by

Chvn I Chu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

R Problem Solving

This document discusses modeling Markov chains in R. It provides examples of creating a transition matrix in R and using packages like "diagram" and "Markovchain" to plot the state space and calculate properties of the Markov chain. It then shows an example of using a transition matrix to find the expected number of coin flips needed to get three heads in a row, modeling it as a Markov chain with states for the number of consecutive heads seen so far.

Uploaded by

Chvn I Chu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Problem

Problem Solving
Solving

With

PRESENTED BY

Asmaa Hamdy 20201032790


Ebtssam Hassan 20201491232
Ziad Ahmed 20201376238
Rahma Ezzat 20201381362
Abdelrahman salah 20201444840
Aya Mohammed 20201379022
Omar Mekkawy 20201375851
1. Box A contains 2 white balls and box B contains 4 red balls at time n,
select one ball from each box and interchange them, considering
𝑋_𝑛 is the number of red balls in A, The transition matrix:

Answer :
Let’s model this Markov Chain using R.
We will start by creating a transition matrix

Code:

Output:

In the above code,


states refer to the state space of the Markov Chain; while P represents
the transition matrix that gives the probabilities of moving from one
state to another.
There is a package in R ‘ Markov-chain’ which can help us save
time in implementing Markov Chains in R.

Code:

Output:

Now, to plot the above transition matrix we can use the R package,
“diagram”.The “diagram” package has a function called
“plot at” that can help us plot a state space diagram of the
transition matrix in an easy-to-understand manner.

Code:

Box.prob length/width ratio of label box


self.shiftx relative shift of self-arrow, in x-direction
self.shifty relative shift of self-arrow, in y-direction

Output:

Now, the above Markov Chain can be used to answer some of the future state
questions. Like P(𝑥_2 =1 | 𝑥_1 = 0) = 1
We can similarly calculate for two steps. Let’s calculate the
probability of coming back to the 0 state in the second step.

Code: markov.p^2

Output:

We have a probability of 0.0625 of returning to state 0 in the second.

The Markov Chain reaches an equilibrium called a stationary state. In


this case, the starting point becomes completely irrelevant. The
stationary state can be calculated using some linear algebra
methods; however, we have a direct function, ‘steady states’, in R,
which makes our lives easier.
Code:

In the stationary state, we have a probability of 0.06666667 of ending


up in the state 0 and a probability of 0.4 of reaching state 2.
Let’s say we have to complete 20 steps. Will 20 steps take our
Markov Chain to the stationary state?

Since the probabilities in the 20th and 21st steps are coming to be
equal, we can say that we have reached the stationary state.

2. A coin is flipped repeatedly until three heads appear in a row. what is


the expected number of flips needed?
States are {No head, 𝐻, 𝐻𝐻, 𝐻𝐻𝐻}
The transition matrix:

Answer :
Let’s model this Markov Chain using R.
We will start by creating a transition matrix of the zone movement
probabilities.

Code:

Output:
We called the “Markov chain” function

Now, to plot the above transition matrix we can use the R package,
“diagram”.The “diagram” package has a function called
“plotmat” that can help us plot a state space diagram of the
transition matrix in an easy-to-understand manner.

Code:

Output:

Now, the above Markov Chain can be used to answer some


of the future state questions. Like P(𝑥_2=H | 𝑥_1 = HH) = 0
Now, how can we find the
expected numbers of flips needed to reach the absorbing state (HHH)?
Simply we can use the Q matrix
it's a subset of the transition matrix X and it represents the probability
of transitions between transient states and itself
〖
We can do our task by the rule (𝐼−𝑄) ^(−1) 〗
Let's do this with R
We used the library MatLab to inverse our matrix by the 'solve'
function

Code:

Output:

We can conclude that ( 8+4+2) = 14


is the expected number of absorbing starting from state 'T'

You might also like