Підтримка
www.wikidata.uk-ua.nina.az
Dostatnya statistika dlya parametra 8 8 displaystyle theta in Theta sho viznachaye deyake simejstvo F8 displaystyle F theta rozpodiliv jmovirnosti statistika T T X displaystyle T mathrm T X taka sho umovna imovirnist vibirki X X1 X2 Xn displaystyle X X 1 X 2 ldots X n pri danomu znachenni T X displaystyle mathrm T X ne zalezhit vid parametra 8 displaystyle theta Tobto vikonuyetsya rivnist P X X T X t 8 P X X T X t displaystyle mathbb P X in bar X mathrm T X t theta mathbb P X in bar X mathrm T X t Dostatnya statistika T X displaystyle mathrm T X takim chinom mistit u sobi vsyu informaciyu pro parametr 8 displaystyle theta sho mozhe buti oderzhana na osnovi vibirki X Tomu ponyattya dostatnoyi statistiki shiroko vikoristovuyetsya v teoriyi ocinki parametriv Najprostishoyu dostatnoyu statistikoyu ye sama vibirka T X X displaystyle mathrm T X X prote spravdi vazhlivimi ye vipadki koli velichina dostatnoyi statistiki znachno mensha vid velichini vibirki zokrema koli dostatnya statistika virazhayetsya lishe kilkoma chislami Dostatnya statistika S S X displaystyle S mathrm S X nazivayetsya minimalnoyu dostatnoyu yaksho dlya kozhnoyi dostatnoyi statistiki T isnuye nevipadkova vimirna funkciya g sho S X g T X displaystyle S X g T X majzhe napevno Teorema faktorizaciyiTeorema faktorizaciyi daye sposib praktichnogo znahodzhennya dostatnoyi statistiki dlya rozpodilu jmovirnosti Vona daye dostatni i neobhidni umovi dostatnosti statistiki i tverdzhennya teoremi inodi vikoristovuyetsya yak oznachennya Nehaj T X displaystyle mathrm T X deyaka statistika a f8 x displaystyle f theta x umovna funkciya shilnosti chi funkciya jmovirnostej zalezhno vid vidu rozpodilu dlya vektora sposterezhen X Todi T X displaystyle mathrm T X ye dostatnoyu statistikoyu dlya parametra 8 8 displaystyle theta in Theta yaksho i tilki yaksho isnuyut taki vimirni funkciyi h i g sho mozhna zapisati f8 x h x g 8 T x displaystyle f theta x h x g theta mathrm T x Dovedennya Nizhche podano dovedennya dlya chastkovogo vipadku koli rozpodil jmovirnostej ye diskretnim Todi f8 x P X x 8 displaystyle f theta x mathbb P X x theta funkciya jmovirnostej Nehaj dana funkciya maye faktorizaciyu yak u tverdzhenni teoremi i T x t displaystyle mathrm T x t Todi mayemo P X x T X t 8 P X x 8 P T X t 8 h x g 8 T x x T x th x g 8 T x h x g 8 t x T x th x g 8 t h x x T x th x displaystyle begin aligned mathbb P X x mathrm T X t theta amp frac mathbb P X x theta mathbb P mathrm T X t theta amp frac h x g theta mathrm T x sum x mathrm T x t h x g theta mathrm T x amp frac h x g theta t sum x mathrm T x t h x g theta t amp frac h x sum x mathrm T x t h x end aligned Zvidsi bachimo sho umovna jmovirnist vektora X pri zadanomu znachenni statistiki T X displaystyle mathrm T X ne zalezhit vid parametra i vidpovidno T X displaystyle mathrm T X dostatnya statistika Navpaki mozhemo zapisati P X x 8 P X x T X t 8 P T X t 8 displaystyle mathbb P X x theta mathbb P X x mathrm T X t theta cdot mathbb P mathrm T X t theta Z poperednogo mayemo sho pershij mnozhnik pravoyi storoni ne zalezhit vid parametra 8 displaystyle theta i jogo mozhna vzyati za funkciyu h x z tverdzhennya teoremi Drugij mnozhnik ye funkciyeyu vid 8 displaystyle theta i T X displaystyle mathrm T X i jogo mozhna vzyati za funkciyu g 8 T x displaystyle g theta mathrm T x Takim chinom oderzhano neobhidnij rozklad sho zavershuye dovedennya teoremi PrikladiRozpodil Bernulli Nehaj X1 X2 Xn displaystyle X 1 X 2 ldots X n poslidovnist vipadkovih velichin sho rivni 1 z imovirnistyu p i rivni 0 z imovirnistyu 1 p tobto mayut rozpodil Bernulli Todi P x1 xn p p xi 1 p n xi pT x 1 p n T x displaystyle mathbb P x 1 ldots x n p p sum x i 1 p n sum x i p mathrm T x 1 p n mathrm T x yaksho vzyati T X X1 Xn displaystyle mathrm T X X 1 ldots X n Todi dana statistika ye dostatnoyu zgidno z teoremoyu faktorizaciyi yaksho poznachiti g p T x1 xn pT x1 xn 1 p n T x1 xn displaystyle g p mathrm T x 1 ldots x n p mathrm T x 1 ldots x n 1 p n mathrm T x 1 ldots x n h x1 xn 1 displaystyle h x 1 ldots x n 1 Rozpodil Puassona Nehaj X1 X2 Xn displaystyle X 1 X 2 ldots X n poslidovnist vipadkovih velichin z rozpodilom Puassona Todi P x1 xn l e llx1x1 e llx2x2 e llxnxn e nll x1 x2 xn 1x1 x2 xn e nllT x 1x1 x2 xn displaystyle mathbb P x 1 ldots x n lambda e lambda lambda x 1 over x 1 cdot e lambda lambda x 2 over x 2 cdots e lambda lambda x n over x n e n lambda lambda x 1 x 2 cdots x n cdot 1 over x 1 x 2 cdots x n e n lambda lambda mathrm T x cdot 1 over x 1 x 2 cdots x n de T X X1 Xn displaystyle mathrm T X X 1 ldots X n Dana statistika ye dostatnoyu zgidno z teoremoyu faktorizaciyi yaksho poznachiti g p T x1 xn e nllT x displaystyle g p mathrm T x 1 ldots x n e n lambda lambda mathrm T x h x1 xn 1x1 x2 xn displaystyle h x 1 ldots x n 1 over x 1 x 2 cdots x n Rivnomirnij rozpodil Nehaj X1 X2 Xn displaystyle X 1 X 2 ldots X n poslidovnist rivnomirno rozpodilenih vipadkovih velichin X1 X2 Xn U a b displaystyle X 1 X 2 ldots X n U a b Dlya cogo vipadku P x1 xn l b a n1 a min1 i nXi 1 max1 i nXi b displaystyle mathbb P x 1 ldots x n lambda left b a right n mathbf 1 a leq min 1 leq i leq n X i mathbf 1 max 1 leq i leq n X i leq b Zvidsi viplivaye sho statistika T X min1 i nXi max1 i nXi displaystyle T X left min 1 leq i leq n X i max 1 leq i leq n X i right ye dostatnoyu Normalnij rozpodil Dlya vipadkovih velichin X1 X2 Xn displaystyle X 1 X 2 ldots X n z normalnim rozpodilom N m s2 displaystyle mathcal N mu sigma 2 dostatnoyu statistikoyu bude T X i 1nXi i 1nXi2 displaystyle mathrm T X left sum i 1 n X i sum i 1 n X i 2 right VlastivostiDlya dostatnoyi statistiki T ta biyektivnogo vidobrazhennya ϕ displaystyle phi statistika ϕ T displaystyle phi T tezh ye dostatnoyu Yaksho d X displaystyle delta X statistichna ocinka deyakogo parametra 8 displaystyle theta T X displaystyle mathrm T X deyaka dostatnya statistika i d1 X E d X T X displaystyle delta 1 X textrm E delta X T X to d1 X displaystyle delta 1 X ye krashoyu ocinkoyu parametra v sensi serednokvadratichnogo vidhilennya tobto vikonuyetsya nerivnistE d1 X ϑ 2 E d X ϑ 2 displaystyle textrm E delta 1 X vartheta 2 leq textrm E delta X vartheta 2 prichomu rivnist dosyagayetsya lishe koli d displaystyle delta ye vimirnoyu funkciyeyu vid T Teorema Rao Blekvela Z poperednogo oderzhuyetsya sho ocinka mozhe buti optimalnoyu v sensi serednokvadratichnogo vidhilennya lishe koli vona ye vimirnoyu funkciyeyu minimalnoyi dostatnoyi statistiki Yaksho statistika T T X displaystyle T mathrm T X ye dostatnoyu i povnoyu tobto z togo sho E8 g T X 0 8 8 displaystyle E theta g T X 0 forall theta in Theta viplivaye sho P8 g T X 0 1 8 8 displaystyle P theta g T X 0 1 forall theta in Theta to dovilna vimirna funkciya vid neyi ye optimalnoyu ocinkoyu svogo matematichnogo spodivannya Div takozhStatistichna ocinka Parametr Teorema Rao BlekvelaDzherelaKartashov M V Imovirnist procesi statistika Kiyiv VPC Kiyivskij universitet 2007 504 s Gihman I I Skorohod A V Yadrenko M V Teoriya veroyatnostej i matematicheskaya statistika Kiyiv Visha shkola 1988 436 s ros Lehmann E L Casella G 1998 Theory of Point Estimation 2nd ed Springer Chapter 4 ISBN 0 387 98502 6, Вікіпедія, Українська, Україна, книга, книги, бібліотека, стаття, читати, завантажити, безкоштовно, безкоштовно завантажити, mp3, відео, mp4, 3gp, jpg, jpeg, gif, png, малюнок, музика, пісня, фільм, книга, гра, ігри, мобільний, телефон, android, ios, apple, мобільний телефон, samsung, iphone, xiomi, xiaomi, redmi, honor, oppo, nokia, sonya, mi, ПК, web, Інтернет
Топ