n ∑ i= 1 ( X − X) i σ 2 Helmert does not assign a name to the distributi<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> the sum <str<strong>on</strong>g>of</str<strong>on</strong>g> squares [λλ]. His objective is to use the result to estimate the probable error. Also <str<strong>on</strong>g>of</str<strong>on</strong>g> interest in the article is the estimati<strong>on</strong> related to the precisi<strong>on</strong> h (and there<strong>by</strong> σ) in a maximum likelihood manner in secti<strong>on</strong> 2, such that: 38 2 2 2h n χ 1 [ λλ] = σ = −1 2 n−1
Review 1 <str<strong>on</strong>g>of</str<strong>on</strong>g> R.A. Fisher’s Inverse Probability 1. Biographical Notes R<strong>on</strong>ald Aylmer Fisher was born in L<strong>on</strong>d<strong>on</strong> in 1890. He received scholarships to study mathematics, statistical mechanics and quantum theory at Cambridge University, where he also studied evoluti<strong>on</strong>ary theory and biometrics. After graduati<strong>on</strong>, he worked for an investment company and taught mathematics and physics at public schools from 1915 to 1919. From 1919 to 1943, he was associated with the Rothamsted (Agricultural) Experimental Stati<strong>on</strong>, c<strong>on</strong>tributing to experimental design theory and the development <str<strong>on</strong>g>of</str<strong>on</strong>g> a Statistics <str<strong>on</strong>g>De</str<strong>on</strong>g>partment. In 1943 he became pr<str<strong>on</strong>g>of</str<strong>on</strong>g>essor <str<strong>on</strong>g>of</str<strong>on</strong>g> genetics at Cambridge, remaining there until his retirement in 1957. He died at Adelaide in Australia, in 1962. 2. Review <str<strong>on</strong>g>of</str<strong>on</strong>g> Inverse Probability In this article, published in 1930, R.A. Fisher cauti<strong>on</strong>s against the applicati<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> prior probability densities for parameter estimati<strong>on</strong> using inverse probability 2 , when a priori knowledge <str<strong>on</strong>g>of</str<strong>on</strong>g> the distributi<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> the parameters is not available (e.g., from known frequency distributi<strong>on</strong>s). Fisher indicates in the first paragraph, that the subject had been c<strong>on</strong>troversial for some time, suggesting that: “Bayes, who seems to have first attempted to apply the noti<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> probability, not <strong>on</strong>ly to effects in relati<strong>on</strong> to their causes but also to causes in relati<strong>on</strong> to their effects, invented a theory 3 , and evidently doubted its soundness, for he did not publish it during his life.” Fisher describes the manner in which a (known) prior density can be used in the calculati<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> probabilities: “Suppose that we know that a populati<strong>on</strong> from which our observati<strong>on</strong>s were drawn had itself been drawn at random from a super-populati<strong>on</strong>…that the probability that θ1, θ2, θ3,... shall lie in any defined infinitesimal range 1 2 3 ... dθ dθ dθ is given <strong>by</strong> dF =Ψ( θ , θ , θ ,...) dθ dθ dθ ..., 1 2 3 1 2 3 then the probability <str<strong>on</strong>g>of</str<strong>on</strong>g> successive events (a) drawing from the super-populati<strong>on</strong> a populati<strong>on</strong> with parameters having the particular values θ1, θ2, θ 3,... and (b) drawing from such a populati<strong>on</strong> the sample values 1 ,..., x x n , will have a joint probability 1 Submitted for STA 4000H under the directi<strong>on</strong> <str<strong>on</strong>g>of</str<strong>on</strong>g> Pr<str<strong>on</strong>g>of</str<strong>on</strong>g>essor Jeffrey Rosenthal. 2 According to H.A. David and A.W.F. Edwards in Annotated Readings in the History <str<strong>on</strong>g>of</str<strong>on</strong>g> Statistics, Springer-Verlag 2001, page 189: “…“Inverse probability”…would not necessarily have been taken to refer exclusively to the Bayesian method (which in the paper Fisher calls “inverse probability strictly speaking”) but to the general problem <str<strong>on</strong>g>of</str<strong>on</strong>g> arguing “inversely” from sample to parameter…” 3 In the mid-1700s. 39
- Page 1 and 2: Review 1 of <stron
- Page 3 and 4: E = pP = 2pA = peA [where p refers
- Page 5 and 6: It is interesting that after having
- Page 7 and 8: The top row of the
- Page 9 and 10: In Fermat’s theory, the player al
- Page 11 and 12: Using Pascal’s terminology, the v
- Page 13 and 14: meaning, in the context of<
- Page 15 and 16: Review 1 of Christ
- Page 17 and 18: Propositions 5 and 6 explain the ex
- Page 19 and 20: Review 1 of Dr. Jo
- Page 21 and 22: Review 1 of Pierre
- Page 23 and 24: A table is given showing the expect
- Page 25 and 26: … 0, 1, 2, ..., r-1, r, r+1, ...,
- Page 27 and 28: number of favourab
- Page 29 and 30: “…the Logarithm of</str
- Page 31 and 32: This is essentially what De
- Page 33 and 34: Review 1 of Friedr
- Page 35 and 36: The determinant is then 1 . 1 . …
- Page 37: where again, h is “the precision
- Page 41 and 42: 1 1 a− n−a− 2 2 P( ς|a,n)d