2007N1221

O'Reilly Statistics Hacksɂ

̖{AO'Reilly{ƂōN(2006N)51Ɋsꂽ̂̓{łłB

҂͊VvBO'ReillyPython NbNubN2ł󂵂lłB
͂̃NbNubN^̓\tgEFÂ׋pɍŵAǂ݂₷Af炵{ƋLĂB

āAb߂Statistics HacksɂāB

̖{ O'ReillyHACKV[YAvƂȂΐHwЂЂ̃ICVC肪Ă{Ȃ̂ACɂȂƂ낪B

́AAmazon()Statistics Hacks̔̔y[WɂāAr[Ă̂C

̒S. DavisƂɂ́An߂10y[Wɂłɓvlł΂킩邾낤F2܂܂ĂAƂ̂B

pƁC

# O'ReillyStatistics Hacks̋Lqł
Page 5:
"The mean will be close to some scores and far away from some others,
but if you add up those distances, you get a total that is as small as possible."

# S.Daviŝɑ΂Lq
Wrong. The mean minimizes the sum of SQUARED distances;
the MEDIAN minimizes the sum of the distances. Hand-waving on the next page apologizing
for how complicated the formula for the standard deviation is because
"there are some mathematical complications with summing distances" would suggest to me
that the omission of "squared" on page 5 was not a mere typo or a misguided attempt at simplification.

O'Reilly̏Ђɂ΁Aϒl̘͋aŏɂ̂AƂĂ邪ႤB
ϒl͋̎aŏɂ̂B
ɂẮA͂ł͂Ȃ낤ƂĂB
܂TO`̂iŁAׂȌ͖肻ȂɂȂ낤B

ɎwEĂ͉̂@藝ɂĂStatistics Hacks̋LqɂĂB

Independent(Ɨ)Mutually exclusive(w)ɂĂ܂ɂĂȂHƂwEłB

page 10:
"Additive rule: the probability of any one of several independent events occuring is
the *sum* of each event's probability." [Emphasis on "sum" is the book's, not mine.]

This isn't just plain wrong; it's cringe-inducingly, forehead-slappingly wrong.
The additive rule is for any one of several mutually exclusive events occuring,
and independent pretty much implies *not* mutually exclusive
(the annoying corner cases being those where some events under consideration are completely impossible anyhow).
The rule for one of several independent events occuring is more complicated;
for example, the probability of a fair coin coming up "heads" on
either its first or second flip is obviously not 1,
as the author's statement would ridiculously imply, but rather 3/4.

@藝"CxgNm͂̃CxgɊ܂܂N肤SĂ̏ꍇ̊m̘aɓB"Ƃ́A

wE҂̌ mutually exclusiveƂtgƁA
"݂ɔrIimutually exclusivejȃCxĝǂ炩Nm͂ꂼ̃Cxg̋Nm̘aɓ"
Ƃ̂炵B

ƗƑݔrI(w)͈ʂɊ֌W͂ȂƂɂẮA
"֌WHȂHim̍uiXjj"
O'Reilly Statistics Hacks{͂ȂɌł͂ȂLqĂ[Ȃ̂łB

āAo{ł̐擪ATvPDF (pdf)Ō邱Ƃł̂Ǎ͏C邱ƂȂÂ܂ܖ󂳂ĂB

Vv͖ɌȂƂdȂ̂ŎdȂƎvAO'Reilly()̕ҏWAmazon()2006N11 (NOł) ɎwEĂ̂ɕuĂ̂܂܏oĂ܂̂낤B

ƂƏcOB

vƂC̗ǂtɂ܂ꂿ_ŐTdɎgȂƊ댯ȃc[Ƃ߂́A\NŉxĂA̖{ՂɋC̗ǂ{ɂȂĂCȂ͂ȂB

܂ŏƂĂȂłAɂ͉߂L̂́Av̊b̗͂̂lɂƂĂEntertainmentƂĊy߂{ɂȂĂ̂ȂƂvB
"Excellent explanation of statistics for "normal" people" (Elihu D. Feustel "daringly" H)

ƎvǁAXŗǂŃ`FbNĂ炾낤ȁB
posted by jitterbug at 00:13| Comment(0) | TrackBack(0) | L | |

L

̍L60ȏXVȂuOɕ\Ă܂B

ȉ̂ꂩ̕@Ŕ\ɂ邱Ƃ\łB

EL̓eAҏWȂ
E}CuÓyݒz @yLݒz Au60ԍXVꍇv uL\ȂvɃ`FbNĕۑB

posted by Seesaa uO
×

̍L1NȏVL̓eȂuOɕ\Ă܂B