ecoli

Evidently with that rapt interest you found in his eyes, you didn't listning to much of the debate. And would you have rather he scowl as Obama began to do after about five minutes into the session? 'course if I was getting my ass kicked as Obama was, I'd begin to scoul myself. Or was that an awkward look of amazement to think that someone would challange His Lordship?

Come on, man, look at this smirk:

Romney needs to wipe that smarmy, self-satisfied grin off his face when he's not talking. Also he needs to stop blinking so much. At one point I counted 64 blinks per minute. Who needs to blink that much?!

the explanation was perfect. thanks for that it made things clearer. i would like to have a follow-up question if you don't mind. the different probabilities associated with the transitions are taken from what kind of source? because i've read a couple of sources and i am quite confused because they have a lot of probabilities incorporated with the different transitions of states. i can't quite find out how these probabilities came about. thanks.

That depends on the application. In the above example, the probability of transition can be calculated by the frequency in which one letter follows another in some word dictionary. For biological sequences, there are similar databases of biological sequences with annotations (where HMM are commonly used to predict function/annotation based on sequence). I'm sure there are many other applications I don't know much about.

Thank you for your reply ecoli and I understand [math] A^-1 [/math]; however, I don't understand why it is necessary to do so --- in other words, what is the application of [math] A^-1 [/math]? Does the elementary matrix act as a "key" that unlocks a matrix, like something you might find in cryptography? It helps me to understand the topics in mathematics if I can understand where it might apply (real-world problem). For instance, does it allow us to construct a more detailed model of whatever is being tested?

That's a useful way to think of it. A practical application of matrix inverses are, because there's no definition of matrix division, multiplying by a matrix's inverse accomplishes the same thing. This is useful when you need to solve a system of linear equations.

Here's a problem that can be solved using inverse matrix multiplication:

A group is traveling by train and bus. Bus tickets cost $3 for children and $3.20 for adults. Train tickets cost $3.50 for children and $3.60 for adults. The group spent $118.40 on bus fare and $135.20 on train fare. How many adults and children were in the group?

You could imagine a really large and complex set of equations (which describes problems in engineering, physics, biology, computer graphics, etc) with a set variables you need to solve. Linear algebra provides a framework to do this by defining operations on matrices of numbers. One of those being matrix inversion, which is especially useful when a single solution exists to this set of equations (for example, only 1 combination of adults and children could have spent that combination on bus and train fare).

We've stepped into matrix operations this week and there are a few theorems that I don't quit understand. Now transposing a matrix is fairly straightforward and easy to comprehend, at least with small matrices. The part I'm having trouble with is the invertible matrix. I understand the key component that A^-1 is the inverse of A; however, I'm absolutely lost by the idea -- E_p...E₂, E₁A=I_n and the algebraic transformations of this equation. What is the purpose in finding the elementary matrix and what is a good way to visualize and understand this concept?

You can think of matrix inversion as analogous to the inversion of any real number.

Just like: [math] 4 * 4^{-1} = 4 * \frac{1}{4} = 1 [/math]

For some matrices, A, there exists a so-called inverse such that:

[math] AA^{-1} = I [/math], I being the identity matrix - a square matrix with 1's along the diagonal and 0's everywhere else.

Elementary matrices allow you to perform operations on matrix A, for example, to solve for the inverse matrix.

Windows opening on airplanes? Maybe if designed by DC engineers.

Not really my field but...

If you have two variables for example height and weight and you measure lots of people and calculate the correlation coefficient and find that, for example, it's 0.8 the you can find R^2 easily enough: 0.64 .

That tells you that 64% of the variation in weight is "due to" the variation in height. The rest of the variability must be due to other factors.

Couldn't you just as easily state the reverse? Since you haven't controlled for weight or height.

shameless bump

For example, I know that r^2 of linear regression is, in certain cases, related to the correlation coefficient (goodness of fit) but is there are precise mathematical definition? Perhaps with a geometric interpretation?

Any and all references would be helpful.

I think elevator scheduling uses similar, online learning approaches.

 

NEW SUB-TOPIC

 

Massive Tumors in Rats Fed Monsanto’s Genetically Engineered Corn

 

The study done by a lab in France.

 

Interestingly the most GMO maize fed to the rats was 33%.

 

Monsanto spokesman Thomas Helscher said the company would review the study thoroughly. However, he added: "Numerous peer-reviewed scientific studies performed on biotech crops to date, including more than a hundred feeding studies, have continuously confirmed their safety, as reflected in the respective safety assessments by regulatory authorities around the world."

 

The statistics on this study are particularly bad, which is getting some coverage in the blogosphere. With 9 treatment groups and 1 control group being the most obvious 'no-no' since over half of the mice in the control group developed tumors anyway.

the problem of multiple comparisons: http://xkcd.com/882/

analysis of this study: http://michaelgrayer.posterous.com/in-which-i-blow-a-gasket-and-get-very-uppity

should be O(n log(n))

It depends on the lab, but everyone starts from nothing and that should be recognized by your post-doc. Just make sure you're looking out for yourself. If your not sure of something, of the calculation of a dilution, etc, just ask. I'm sure the post-doc would rather be bothered with questions than a communal reagent gets made incorrectly.

pure placebo effect: http://www.sciencedaily.com/releases/2011/03/110324104147.htm

There is no evidence for chi lines, and other such things.

hi, can i ask for a little assistance in understanding HMM? i've been reading and watching tutorials about it but none seemed to have discussed it as to my understanding. i am a fourth year university student trying to have a better understanding of HMM. any assistance would be of help. thanks.

 

p.s. also i want to know if bayesian is really needed for HMM to work.

You don't need to construe HMM in a bayesian framework, though it is statistical.

The example I like to think of is text prediction, especially for example, when sending a text message from a cell phone (autocomplete). The problem is: can you predict the complete word the author intends from a partial, incomplete word.

So you have a partial string, something like: "goodb". the model treats each letter as a node and the transition/edge between the nodes represents the probability that that letter followed from the previous one. This can be calculated from the frequency of letter distribution corpus/dictionary of commonly used words. For example the probability that the character 'u' follows 'q' in standard american english is quite high. two weird consonants 'jz' probably less so.

So the user types in "goodb" what is the most probable word? the transition path with the highest probability would [i assume] be "-y-e" followed by an "end" state.

Did this help or is this too basic for you?

I should say I'm referring to the name as used for men.

For men, yeah, though I wouldn't say it's very common.

but a semi-famous example: http://en.wikipedia.org/wiki/Sidney_Beckerman_(musician)

No.

 

I wish people understood how a risk assessment works, because it's a perfectly reasonable and simple approach how to deal with any risk, in any situation - also if you have many unknowns.

 

Risk assessment certainly does not advise you to assume the worst, unless you have the intention of going bankrupt trying to prevent the most unlikely scenarios.

Risk assessment tells you to look at every scenario, and look at the effects and the chance it happens.

 

Risk = effect * chance

 

For example, if a nuclear holocaust happens it's very very very bad. But the chance it happens is really small (please don't believe the scaremongering). The risk is then reasonable.

The effects of some local terrorism in Israel is much smaller, after all only a few people will be directly affected. But the chance that will happen is nearly 1. It's almost a certainty. So, that risk is a bigger risk than Iran attacking with a nuke.

 

Iranians are not completely stupid. They know that if they lob a nuke on Israel, their own country will be reduced to a molten sea of sand too. They might use strong words, but that's a lot of bluff, which sounds good on Iranian TV. Also, for all we know, they might have a new and more reasonable president soon. Ahmadinejad might actually lose the elections... they do have a functioning democracy in Iran. This year they elect a new parliament, and next year a new president. You might argue that they don't seem to have a lot of choice, but when I look at the current US elections, I get that same impression

Nuclear weapons are an existential risk, however, so the effect always outweighs the probability, no matter how small particularly in light of zapatos' post.

Isn't it safer for Israel to assume the worst if Iran is hampering the efforts of UN nuclear inspectors?

Both of them ducked the vaccine question. Obama pretending that the question was about affordability of vaccines (which is obviously not the problem when it comes to vaccine denialism of liberal suburbanites) and Romney answered a question about vaccine development - an important issue but not what was asked.

Romney wins the word count contest.

Then again... If I had to visit one it would almost have to be Dark Angel because, you know, genetically engineered Jessica Alba.

+1 underrated show, if you ask me.

I agree with the editors note... "While they have way too much in common, the actual Taliban uses political violence to achieve its ends and the Tea Party doesn’t — and that’s an important distinction."

The Boy scouts of america and the Hitler Youth were two organizations that 1) made up of young people 2) officially shunned homosexuals... but that doesn't make any comparison between the two in any way appropriate or useful. Unless of your goal isn't to make an apt comparison but to the poison the well by guilty association.

This one speaks for itself.

 

 

 

What do you think? Agree? Disagree? Any ideas to improve the situation?

A ridiculous analogy perpetrated, though not originated, by Alan Sorkin. For example, when have tea party members ever committed war crimes: http://www.slate.com/articles/news_and_politics/fighting_words/2011/02/news_flash_the_taliban_violate_human_rights.html

you haven't really asked a question...

RFID chips are proof of god? Sorry not good enoguh.