Reader Comments (rss)

(Flattened / Threaded)

1.  

   Um Sony?

   

   Sissy Pants, Nov 18th, 2005 @ 9:53am

   
   

   Well, considering the Sony rootkit has been "infecting" computers since 2004 the extent isn't known yet... it's not a huge suprise that 80% of china is infected...
   
   What percentage of the US is infected? Or Europe? Are you?

   [ reply to this | link to this | view in thread ]

2.  

   Re: Um Sony?

   

   Testudo, Nov 18th, 2005 @ 9:59am

   
   

   I agree. Some viruses may not become appearent for a long time. Not all viruses are the headline grabbers that bring down entire computer networks within days

   [ reply to this | link to this | view in thread ]

3.  

   Confidence Interval?

   

   dorpus, Nov 18th, 2005 @ 10:32am

   
   

   I would want to know the 95% CI on that. There is the concept of Risk Function -- mathematically, it is very difficult to prove whether a coin flip has a 50% chance of heads, or a 55% chance of heads.
   
   To determine probability, our intuition may tell us to flip 100 times and count the number of heads, i.e. use the sample proportion. However, it can be shown mathematically that it is better to artificially add 3 heads and 3 tails to our counts, if we are confident that the true probability is close to 50%. Our estimate will be about 11% more accurate this way.
   
   
   

   [ reply to this | link to this | view in thread ]

4.  

   Re: Confidence Interval?

   

   dorpus, Nov 18th, 2005 @ 10:39am

   
   

   p.s. When humans flip a coin, they tend to place the head up before flipping, and everyone has roughly the same thumb strength, so a paper from Stanford last year showed that the probability of a coin landing heads is in fact roughly 55%.
   
   

   [ reply to this | link to this | view in thread ]

5.  

   Re: Confidence Interval?

   

   NormL, Nov 18th, 2005 @ 11:11am

   
   

   Why would you want to know? Why don't you want to know at this point in time? What is causing this delay in your decision to gain the knowledge in which you seek?

   [ reply to this | link to this | view in thread ]

6.  

   Re: Confidence Interval?

   

   Wyatt, Nov 18th, 2005 @ 11:13am

   
   

   Why not the 75% conf. interval? What's so special about 95%? Would you not make a bet on a study that had had a 90% CI, what about 85%? Stats is stats... not truth.

   [ reply to this | link to this | view in thread ]

7.  

   Preventing viruses requires education

   

   Duodave, Nov 18th, 2005 @ 12:19pm

   
   

   Seeing as the Chinese government has done everything in its power to prevent their citizens from communicating with the free world, it should come as no surprise they are infected. They obviously have no clue how to prevent them. The great wall of virii.

   [ reply to this | link to this | view in thread ]

8.  

   Re: Confidence Interval?

   

   antigone, Nov 18th, 2005 @ 12:25pm

   
   

   By using a higher confidence interval, you're less likely to make a Type II error, which is a really stupid thing to make - to claim something is statistically significant when it isn't. It's like the commercial for the prescription for Detrol LA. When you actually look at the reports for it, the medication really didn't do anything that was significant.
   
   And the 95% CI is a typical cutoff for estimating unusual values.

   [ reply to this | link to this | view in thread ]

9.  

   hen Only 80% Of Computers Having Viruses Is Consid

   

   Bryan H, Nov 18th, 2005 @ 1:21pm

   
   

   It's probably China's government infecting the computer's.
   
   This way the Chinese are so damn busy cleaning their computers of viruses there isn't any time to fight the government.

   [ reply to this | link to this | view in thread ]

10.  

    Re: Confidence Interval?

    

    spiker, Nov 18th, 2005 @ 1:31pm

    
    

    With all due respect I think you have type 1 error mixed up with type 2 error

    [ reply to this | link to this | view in thread ]

11.  

    Re: Confidence Interval?

    

    dorpus, Nov 18th, 2005 @ 2:08pm

    
    

    Actually, that's a Type I error, when you conclude that something is significant (when it isn't). A Type II error will conclude the effect is insignificant, when in fact it is significant.
    
    The probability of not making a Type II error, i.e. correctly finding significance, is known as "power".
    
    95% CI has been the gold standard for much of scientific research for many years. It's saying that if we were to repeat the experiment at the same scale, we have a 95% chance of reaching the same conclusion. Here, the 95% is known as the "size" of the test. One can perform 90% CI in situations where the results are inexact, or 99% CI's in situations that demand more precision.
    
    
    

    [ reply to this | link to this | view in thread ]

Add Your Comment