Statistics and Computer Science
I have to take some statistics courses for my degree and I'm having a hard time staying interested. I think it would help to know that statistics is very important to computer science. Please tell me how and why.
8 comments
Statistics is probably the most useful maths-based subject hat I did badly in (out of disinterest). It wasn't until a year after my statistics course that I realised its importance..
I used statistics in a uni project to analyse audio data to perform event detection. The goal of my part of the project was to determine the most optimal balance between accuracy of event detection and power consumption (it was to be run on embedded hardware). I wrote some tools which gave different statistical data of an audio clip, which we could then use to determine a power-efficient, yet accurate, algorithm to do real time event detection. The program gave such data as root mean square of the audio clip, its skewness, kurtosis, standard deviation and a bunch more, as well as combinations of these (and graphed them visually, superimposed on the sound wave).
That was my first hands-on intro to statistics. Other things I've encountered since then is statistical clustering for the purpose of data classification and pattern matching, statistical analysis of text for spam detection and other such things.
I wish I'd payed more attention in class as I find statistics to be one of the more useful things which I didnt learn.. though I bought an intro book on statistics recently and am working my way through it, hopefully learning what I missed.
The problem with my course was that it wasn't very hands on. It was too abstract, so it was difficult to see the value in it and therefore easy to neglect. The thing is, though, that a lot of technology we use every day relies on statistics. If I can give you any advice, it would be to force yourself to do well in statistics - it'll pay off in the long run.
I used statistics in a uni project to analyse audio data to perform event detection. The goal of my part of the project was to determine the most optimal balance between accuracy of event detection and power consumption (it was to be run on embedded hardware). I wrote some tools which gave different statistical data of an audio clip, which we could then use to determine a power-efficient, yet accurate, algorithm to do real time event detection. The program gave such data as root mean square of the audio clip, its skewness, kurtosis, standard deviation and a bunch more, as well as combinations of these (and graphed them visually, superimposed on the sound wave).
That was my first hands-on intro to statistics. Other things I've encountered since then is statistical clustering for the purpose of data classification and pattern matching, statistical analysis of text for spam detection and other such things.
I wish I'd payed more attention in class as I find statistics to be one of the more useful things which I didnt learn.. though I bought an intro book on statistics recently and am working my way through it, hopefully learning what I missed.
The problem with my course was that it wasn't very hands on. It was too abstract, so it was difficult to see the value in it and therefore easy to neglect. The thing is, though, that a lot of technology we use every day relies on statistics. If I can give you any advice, it would be to force yourself to do well in statistics - it'll pay off in the long run.
This is from the probability course I took a year ago:
"Probability theory has become indispensable in computer science. In areas such as artificial intelligence and computer science theory, probabilistic methods and ideas based on randomization are central. In other areas such as networks and systems, probability is becoming an increasingly useful framework for handling uncertainty and modeling the patterns of data that occur in complex systems." - http://www.cs.cmu.edu/~15359/
Don't lose track of the end goal. A part of earning your degree is to build your toolbox and skillset. Competency in statistics and/or probability theory can be another asset for you after you leave the ivory towers of academia.
(Yes, I understand statistics != probability. In fact, probability theory is a subset of statistics.)
"Probability theory has become indispensable in computer science. In areas such as artificial intelligence and computer science theory, probabilistic methods and ideas based on randomization are central. In other areas such as networks and systems, probability is becoming an increasingly useful framework for handling uncertainty and modeling the patterns of data that occur in complex systems." - http://www.cs.cmu.edu/~15359/
Don't lose track of the end goal. A part of earning your degree is to build your toolbox and skillset. Competency in statistics and/or probability theory can be another asset for you after you leave the ivory towers of academia.
(Yes, I understand statistics != probability. In fact, probability theory is a subset of statistics.)
How about this:
“I keep saying that the sexy job in the next 10 years will be statisticians,” said Hal Varian, chief economist at Google. “And I’m not kidding.”
“I keep saying that the sexy job in the next 10 years will be statisticians,” said Hal Varian, chief economist at Google. “And I’m not kidding.”
Statistics and probability are the foundation of Machine Learning. Information Theory and Machine Learning are practically homomorphic(1,2). You can't really be interested in Computer Science without being interested in Information Theory. Therefore, you're interested in Statistics. QED.
(1)http://themachinelearningforum.org/index.php/overviews/34-co...
(2)http://people.cs.ubc.ca/~murphyk/Teaching/CS340-Fall07/infoT...
(1)http://themachinelearningforum.org/index.php/overviews/34-co...
(2)http://people.cs.ubc.ca/~murphyk/Teaching/CS340-Fall07/infoT...
Statistics has direct applications in:
* VM optimization
* Data Mining
* Machine learning/classification
* Hueristics
* Benchmarking (and other measurment)
Indirectly there are even more. There isn't a week that goes by where I don't go looking for solutions in statistical methods.
* VM optimization
* Data Mining
* Machine learning/classification
* Hueristics
* Benchmarking (and other measurment)
Indirectly there are even more. There isn't a week that goes by where I don't go looking for solutions in statistical methods.
I'm in the same boat, I graduated with a computer degree but I didn't learn much about stats. Since then I've gone back to take part time courses to catch up. I find it really interesting and useful. I wish my school made stats a part of the required curriculum.
Stats is where it's going to be at. And Gauss is your friend. He makes it hard for lotsa certification junkies to flood the market.
A masters in stats is really something.
A masters in stats is really something.
We are entering the age of statistics. Don't miss out.