Comparing Exponential Distribution With Central Limit Theorem

Some notes based on my statistical studies. A good reference on how do simulations in R to prove and gain better grasp in understanding the Central Limit Theorem (CLT), and how CLT can even be used when your original data is not a normal distribution.

Or maybe perhaps you just prefer the video - Understanding Central Limit Theorem - with Bunnies and Dragons

Overview


In this report we will investigate the exponential distribution in R and compare it with the Central Limit Theorem.

Simulations


The exponential distribution will be simulated in R using rexp(n,lambda) where lambda is the rate parameter. The mean of exponential distribution is 1/lambda and the standard deviation is also 1/lambda. We will use lambda = 0.2 for all of the simulations.
We will illustrate via simulation and associated explanatory text the properties of the distribution of the mean of 40 exponentials. Among our findings would be:
  1. The sample mean and it’s comparison to the theoretical mean of the distribution.
  2. The sample variance and it’s comparison to the theoretical variance of the distribution.
  3. The distribution and it’s normality.
Before we begin, let’s jot down some facts:
  • The mean of the exponential distribution = 1/0.2 = 5.
  • The standard deviation of the exponential distribution = 1/0.2 = 5.
  • The sample size = 40.
  • The simulation will be run 1000 times.

Sample Mean versus Theoretical Mean


We start by illustrating the sample mean
# set seed so that our randomly generated graph remains the same for each run.
set.seed(10000)

#generate a sample of size 40 and iterate 1000 times.

mns = NULL
for (i in 1 : 1000) mns = c(mns, mean(rexp(40,rate=0.2)))

hist(mns, main="1000 averages of 40 random uniform", cex.main = 0.7)
abline(v=mean(mns),col="green")
abline(v=5,col="red")

legend( 6, 250, c("Sample Mean","Theoretical Mean"),lty=c(1,1),lwd=c(1,1,1),col=c("green","red"), cex = 0.55)
#find out the mean
mean(mns)
## [1] 5.00599
As can be seen, the mean of the sample is 5.01. This is very close to the theoretical mean, 5.0.

Sample Variance versus Theoretical Variance


Now we study the variability of the sample.
Considering the sample we’ve used previously, let’s find the variance. Or rather, let’s find the standard deviation for the graph since we know that standard deviation is basically just the square root of the variance - thus it should be sufficient in our attempt to prove the variability of the variance.
#get the standard deviation
#sd(mns)
hist(mns, main="1000 averages of 40 random uniform", cex.main = 0.7
     , xlab="Mean Averages")
abline(v=mean(mns),col="green")
abline(v=mean(mns)+sd(mns), col="red", lwd=2.5)
abline(v=mean(mns)-sd(mns), col="red", lwd=2.5)

#now add the theoretical standard deviation  
abline(v=mean(mns)+sqrt(5^2/40), col="blue", lwd=2.5)
abline(v=mean(mns)-sqrt(5^2/40), col="blue", lwd=2.5)

legend( 6, 250, c("Sample Mean","Sample SD","Theoretical SD"),lty=c(1,1),lwd=c(1,1,1),col=c("green","red","blue"), cex = 0.55)
Based on the above the graph, we see that the sample standard deviation is very close to the theoretical standard deviation.
Thus we conclude that the variance is approximately similar.

Distribution and Conclusion


Now let’s compare the distribution of a large collection of random exponentials (let say the size is 1000) against a distribution of a large collection of averages of 40 exponentials (let us say the collection size is 1000).
#generate the large sample of size 1000
large <- span=""> rexp(1000, rate=0.2)

mean(large)
## [1] 4.977901
par(mfrow=c(1,2))    
hist(large, main="1000 random uniform", cex.main = 0.7)
abline(v=mean(large),col="blue")
abline(v=5,col="red")
legend( 6, 250, c("Sample Mean","Theoretical Mean"),lty=c(1,1),lwd=c(1,1,1),col=c("blue","red"), cex = 0.55)

mean(mns)
## [1] 5.00599
hist(mns, main="1000 averages of 40 random uniform", cex.main = 0.7)
abline(v=mean(mns),col="blue")
abline(v=5,col="red")
legend( 6, 250, c("Sample Mean","Theoretical Mean"),lty=c(1,1),lwd=c(1,1,1),col=c("blue","red"), cex = 0.55)
Based on the simulation above, the mean for the first graph turned out to be 4.978, while the mean for the second graph was 5.01. Both values are really close the theoretical mean.
Looking at the histograms we see that the right histogram (average over sample) is more gaussian looking compared to the one the left.
We see that by taking repeated average many times across many small samples, we were able to derive a more accurate mean result.
In conclusion, the central limit theorem stipulates that a computed values of the average will be distributed accordingly to the normal distribution. Since our graph looks like a bell curve - we can claim that our graph approximately obeys the normal distribution.

Comments

Post a Comment

Popular posts from this blog

HIVE: Both Left and Right Aliases Encountered in Join

Recently was stuck on this error while trying to do a JOIN in HIVE. Below is my SQL create table ma.internet_v2 as select A.msisdn, A.imsi, A.dt, A.start_time, A.end_time, A.url, A.ttl_connection_dur_ms, A.ttl_upload_bytes, A.ttl_download_bytes, A.ttl_cdr_cnt, coalesce(B.domain_desc,A.domain_desc) domain_desc, coalesce(B.subdomain_desc, A.subdomain_desc) subdomain_desc from db.internet A left outer join hfz_domain_name_mapping B on instr(A.url, B.url_pattern) > 0; HIVE then give me and error message "Both Left and Right Aliases Encountered in Join". After going through some mail threads it seems that HIVE can support certain types of join. Only equality joins, outer joins, and left semi joins are supported in Hive. Hive does not support join conditions that are not equality conditions as it is very difficult to express such conditions as a map/reduce job. A poster from  Stack Overflow  suggested that one should use WHERE instead of ON. Duly following his advice,
Keep reading

MVNO: Through thick and thin

The below article was taken from Silicon.com. Click here to read the actual article. By Natasha Lomas, 27 October 2008 16:23 MVNO? That'll be an abbreviation - another pointless four letter one no doubt… Right and wrong. MVNO stands for mobile virtual network operator - and it refers to a rather intriguing slice of the mobile pie. I'm listening… Building and operating a mobile network is an expensive business, not least because it requires radio spectrum, so the number of mobile network operators (or MNOs) is limited to, at most, a handful per market. For instance, the UK market has five: 3, O2, Orange, T-Mobile and Vodafone. Spectrum does not come cheap. The auction for 3G licences in the UK - way back in 2000 - ended with five companies shelling out a whopping £22.47bn for the privilege of flogging third-generation services to Brits. So the requirement for large capital expenditure to set up a mobile network goes some way to explaining why actual networks are not 10-a-penny.
Keep reading

Changing Windows Account password using Outlook

http://theessentialexchange.com/blogs/michael/archive/2007/11/24/changing-user-passwords-with-outlook-2003-2007.aspx . 1) Call the helpdesk - the most likely scenario if you are not in a hosted environment 2) Use the control panel - most Hosted Exchange providers provide a function on a user's control panel enabling the user to change their password, using the functionality in item (3). Most large companies will implement a similar capability for their users. Small companies will likely need to implement item (3) directly or their helpdesk will do it (as in item 1). 3) Enable IISADMPWD - use KB 555071 to enable the use of a web page to allow changing the password of an active directory user.
Keep reading