From nobody@digitalkingdom.org Sun Oct 28 01:43:25 2007 Received: with ECARTIS (v1.0.0; list lojban-beginners); Sun, 28 Oct 2007 01:43:28 -0700 (PDT) Received: from nobody by chain.digitalkingdom.org with local (Exim 4.67) (envelope-from ) id 1Im3jc-0007sO-6k for lojban-beginners-real@lojban.org; Sun, 28 Oct 2007 01:43:19 -0700 Received: from py-out-1112.google.com ([64.233.166.182]) by chain.digitalkingdom.org with esmtp (Exim 4.67) (envelope-from ) id 1Im3jA-0007rd-Uq for lojban-beginners@lojban.org; Sun, 28 Oct 2007 01:42:59 -0700 Received: by py-out-1112.google.com with SMTP id a25so2290047pyi for ; Sun, 28 Oct 2007 01:42:21 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=s64Hv8LWc5O/0Jk+nphtE2iQ3qRRU5gtcQaAnTIEOqU=; b=TNzmN0ena7ZjAbW3PqaT+ew+R31LHPoySX0yo5JIQxBEuwZo+fEljNNOS52n9IxrtxJPgD/8dCJKTbKBiPbUF2qv/NdA+wt1jwvdVXWtn8HZIy5pV7LjA00we33d2wpwOFI07oVhFLfMVD0mkW7ni7pSOEbAW6e0wkfT7sxBWXw= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=IGgAHaU+WHSdIVwZ+td4QixGCiQiS6NwWf+UF0kiwu20QhcZoXYbrgi39GYh4nNllyBTkB/82mFpV38LoGxpyjlGCkdX3ODXYukloyesZ1bUQ8EIBlHhFcyfqjYpZYyn26qw96hJNTkkP3Jf+QPVra+geA5VS/JN7L/wyXVUCp0= Received: by 10.65.53.3 with SMTP id f3mr10040393qbk.1193560940914; Sun, 28 Oct 2007 01:42:20 -0700 (PDT) Received: by 10.65.252.4 with HTTP; Sun, 28 Oct 2007 01:42:20 -0700 (PDT) Message-ID: Date: Sun, 28 Oct 2007 01:42:20 -0700 From: "Stephen Pollei" To: lojban-beginners@lojban.org Subject: [lojban-beginners] Re: "With /n/ confidence..." In-Reply-To: <5e124660710190256jbf83da2y7d4e44bcf1ab8283@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by Ecartis Content-Disposition: inline References: <925d17560710180911r61f51d1au6e5c90b63b54a724@mail.gmail.com> <20071018164319.GE14269@digitalkingdom.org> <004701c811ae$c82d68e0$e18e6d0a@Starlight> <925d17560710181039p4f40ff00r7db1a6dd0db79575@mail.gmail.com> <5e124660710190256jbf83da2y7d4e44bcf1ab8283@mail.gmail.com> X-Spam-Score: -0.0 X-Spam-Score-Int: 0 X-Spam-Bar: / X-archive-position: 5718 X-ecartis-version: Ecartis v1.0.0 Sender: lojban-beginners-bounce@lojban.org Errors-to: lojban-beginners-bounce@lojban.org X-original-sender: stephen.pollei@gmail.com Precedence: bulk Reply-to: lojban-beginners@lojban.org X-list: lojban-beginners On 10/19/07, Wim Coenen wrote: > 2007/10/18, Jorge Llambías : > > On 10/18/07, Marjorie Scherf wrote: > > > > > > Is probability the same as certainty? > > > > Usually not. I can be 100% certain that the probablility > > of a coin coming up heads is 50%, for example. I would almost never say that I'm 100% certain about much of anything. I used to have a double headed quarter ; I might still have it somewhere. That I'm certain(100%) that it would come up heads 100% of the time. For real life coins I wouldn't say it's exactly 50/50 , there have been known for slight biases to exist.. I would say for normal non-tampered with coins that the probability of it coming up heads is 49%-51% with confidence level exceeding 90% . http://en.wikipedia.org/wiki/Confidence_interval is probably what David Cortesi was referring to. [[ "Confidence levels are typically given alongside statistics resulting from sampling." . In a statement "we are 90% confident that between 35% and 45% of voters favor Candidate A", 90% is our confidence level and 35%-45% is our confidence interval. ]] I've heard one scientist say that if a research paper gives a probability without a confidence interval then he immediately junks the whole paper. cu'o, cunso, pacna, and kanpe are the likely candidates... however only cunso has any real chance of being scientifically useful. cunso cun cu'o random 'chance' x1 is random/fortuitous/unpredictable under conditions x2, with probability distribution x3 the x3 might have a proper range and confidence value given.. however I haven't seen any good examples of how to do that. I assume distribution implies that it's not just a simple point value. For common use the other forms are probably good enough, because I don't take much stock in people's subjective sense of expressing probabilities in numeric form. It might be helpful if standard/epistemology/method place was added to some of these gismu . Maybe the Bayesian and frequentest methods are different enough that lujvo or fu'ivla for each method with it's own unique place structures should be added. http://www.lojban.org/tiki/tiki-index.php?page=BPFK+gismu+Proposal%3A+kanpe&bl > > Certainty is about knowledge, it is subjective. Probability > > can be objective. epistemology .. by what method do we determine the result and know it. The gismu just cry out for another place IMHO .. >> Maybe nobody knows what the probability > > of something happening is, but we can do some experiments > > and find out. We don't get a simple probability but probability ranges with different confidence intervals... As we collect more data we can be more sure of our results and our probability range will probably change for a given confidence interval. > > The probability of the event doesn't change as > > we perform the experiments, If *ALL* the relevant conditions still hold.. cunso and lakne both properly include a "condition(s)" place. > > but our certainty about whether > > it will happen or not might change depending on how accurately > > we are able to determine its probability. The probability range for a given confidence interval will hopefully shrink in size. Of course maybe later data will show that the earlier data didn't give a good picture and the range could in theory grow. > If we are talking about Baysian probability, then there is no > distinction between certainty and probability. Very little. > The distinction only > exists in the "frequency probability" interpretation of probability. > At least I think so. I'm not certain ;) Sure but in Bayesian I'd like to know what your "prior probability" was. I actually would like to have you express the outcome with maybe at least three "prior probabilities" to see if you have enough evidence to cause the results to converge . For normal cases 0.001,0.5, and 0.999 might work well.. For something like a court case 0.0001,0.49,0.51, and 0.9999 at least. If you start with a suspect from New York then 1/8,214,426 or about 0.0000001 might be a good place to start... see if you have enough Bayesian evidence to raise it beyond the standard of proof required(the balance of probabilities (BOP), called the preponderance of evidence, beyond a reasonable doubt, clear and convincing evidence, Air of reality, Reasonable suspicion, Probable Cause) . Then you might notice that the legal standards rarely give objective number standards, but warm and fuzzy subjective descriptions. > http://en.wikipedia.org/wiki/Baysian is an interesting read. Sure here are a few more. http://en.wikipedia.org/wiki/Credible_interval http://en.wikipedia.org/wiki/Bayesian_inference http://en.wikipedia.org/wiki/Fuzzy_logic http://en.wikipedia.org/wiki/Degree_of_truth http://en.wikipedia.org/wiki/Frequentism http://en.wikipedia.org/wiki/Bootstrapping_%28statistics%29 http://en.wikipedia.org/wiki/Burden_of_proof http://www.fooledbyrandomness.com/ Nassim Nicholas Taleb The Black Swan: The Impact of the Highly Improbable, 2007. Fooled by Randomness: The Hidden Role of Chance in the Markets and Life Doubt and Certainty by Tony Rothman and George Sudarshan .