• Nie Znaleziono Wyników

Tests of Univariate Normality

N/A
N/A
Protected

Academic year: 2021

Share "Tests of Univariate Normality"

Copied!
19
0
0

Pełen tekst

(1)

A C T A U N I V E R S I T A T I S L 0 .0 2 I E N S 1 S FO LIA OECONOMICA 40, 19R5

W i.asiaw Wagner *

T EST S OF UNIVARIA TE NORMALITY /

l. _ _ I n t r o d u c

1

1

0

n

A la r g e c l a s s o f goodness o f f i t t e s t s in th e th e o r y of s t a t i s t i c a l in fo r e n c e a re t e s t s o f u n i v a r i a t e n o r m a lity . T e s ts o f n o r m a lit y e n a b le t e s t in g the goodness o f f i t o f a sampled d i s ­ t r i b u t i o n f u n c t io n by th e norm al d i s t r i b u t i o n f u n c t io n of a g iv - en randon v a r i a b l e .

H i s t o r i c a l l y , th e t e s t s and b2 * b u i l t in the t h i r t * l e a , a re the f i r s t t e s t s f o r n o r m a lit y . O th e r t e s t s w ere d e ve ­ loped by G e a ry ( t e s t g ) , O a v id e t a l , ( t e s t n ) , Kolm ogorov ( t e s t 0 ) , Cram er and von M ise s ( t o s t C M ), Anderson and D a r lin g ( t e s t A2 ) , S h a p ir o and W l.lk ( t e s t W ), O 'A g o s tin o ( t e s t Y ) and o t h e r s . B e s id e the above m entioned t e s t s t h e r e a re « a n y m o d if ic a t io n e w hich a r e w id e ly d le c u e s e d in th e p a p e r .

L e t a random v a r i a b l e X o f a c o n tin u o u s typ e be d i s t r i b u t e d In the **ay d e te rm in e d by a d i s t r i b u t i o n f u n c t io n F ( x ) and w it h d i s t r i b u t i o n p a ra m e te rs ji • E ( X ) and cf2 ■ D2 (X ).T h e f a c t t h a t the v a r i a b l e X i s n o rm a lly d i s t r i b u t e d w it h p a ra m e te rs and d I s denoted as X - N ( |j , d2 ) . L e t the sequence X j , . . . , xn ( { x j , d enote a sam ple c o n s is t in g o f n in d ep e n d en t o b s e r v a t io n of v a ­ r i a b l e X, and x and a r i t h m e t i c mean, S 2 - the sum o f sq uarad d e v ia t io n s , , s2 - v a r i a n c e , and s - s ta n d a rd d e v ia t i o n from the sam p le. Sam ple { y j d e n o te s n o n - d e c re e s in g o rd e re d o b s e rv a ­ t io n s o f

3

ample { XA}» 60 th a t y^, < . . . £ y n * F u n c t io n pn^x ^

* L e c t u r e r a t th e D epartm ent o f M a th e m a tic a l and S t a t i s t i c a l M othods, Academy o f A g r i c u l t u r e , Poznart.

(2)

13 an e m p ir ic a l d i s t r i b u t i o n f u n c t io n , v.hare x I s an a r b i t r a r y r o a l number, I . e . tho o b s e r v a t io n f u n c t io n « x , w h llo $ ( u ) nnd $ ~ * ( p ) a re th e norm al d i n t r i b u t i o n f u n c t io n and q u a n t llo o f the p -tb o r d e r of d i s t r i b u t i o n M ( 0 , 1 ) , r e s p e c t i v e l y . F o r a d e te rm in e d x the v a lu e o f F n ( x ) i s a random v a r i a b l e 0 , 1 ( n , 2 ) n I 0 0 0 $ 1 •

The o r d o r s t a t i s t i c i s a random v a r i a b l e b eing tho k - th v a r i a b l e in tho sample { y ^ } . The aomple ( y ^ j i s c a l l e d tho o r d e r s t a t i s t i c . Whon X ~ N ( y , cf2 ) than y^ I s a norm al o r ­ d e r s t a t i s t i c , { y A J- - a sequence o f norm al o r d e r s t a t i s t i c s , and ■|ui } a sequence o f fJ(0 , 1) - o r d e r o t a t l o t l c s , where u^ » ■ ( y 4 - f O / d ). Tha o rd o r s t a t i s t i c u^ has d i s t r i b u t i o n parame- t e r s j E ( u i ) - mi# D2 ( u i ) * v ^ , C o v (u 1# u^) ■ v ^ ; 1, J -* 1 , . . . . n, then f o r N ( ¿j f d ) - o rd o r s t a t i s t i c s we have E i y j ) ■ e ♦ d u t . 0 2 ( y i ) - and C o v (y 1 , y ) - ^ ^ i j * The v a lu e s o f n i "* wi n > 1 “ 1 * •••» [ n/ 2 ] a re t a b a la r lz a d f o r v a r i o u s n s , w h ile v ^ ■ v ^ n and v ^ ■ v^^ n f o r i , J ■ 1 ,..., [ n / 2 ] ; i i J o n ly f o r n » 1, . . . , 20 ( c f . ^ 2 0 ], t a b le s 9 and 1 0 ). 2. The H yp o th es i s o f Goodness o f F i t o f th e E m p ir ic a l D i s t r ib u t io n w it h th e Normal D i s t r i b u t i o n

The s t a t i s t i c a l h y p o th e s is o f the e m p ir ic a l and norm al d i ­ s t r i b u t i o n s i s fo rm u la te d as f o llo w s . L e t F n d en o te a c la s s o f norm al d i s t r i b u t i o n f u n c t io n and G - a c la s s o f d i s t r i b u ­ t io n f u n c t io n o f random v a r i a b l e s h a v in g the t h i r d moment ( P

3

) o t h e r than z e ro and f i n i t e fo u r t h c e n t r a l moment (¿*

4

)# a t fl n G « 0 . L a t u

3

put a h y p o th e s is H^ : F e F^ and H^ i F e G, t h a t th e f u n c t io n F ( x ) b e lo n g s to the c l a s s o f d i s t r i b u t i o n f u n c t io n s and G. Tho hyp

0

t h

060

£> HQ o r Hj w i l l be sim p le i f the d i s t r i b u t i o n s b e lo n g in g to c la s s e s F^ o r G hava th e known d i s t r i b u t i o n p a ra m e te r s .

The h y p o th e s is HQ a g a in 3 t H^ i s v e r i f i e d u s in g one o f the t e s t s f o r n o r m a lit y . G e n e r a l l y , th e y a re d iv id e d , a c c o r d in g to th a s t r u c t u r e o f t a s t s t a t i s t i c s In t o

(3)

1 ) P e a r s o n 's X t e a t s fo r goodness o f f i t ,

2 ) t e s t s based on the com parison o f e m p ir ic a l and nortnnj. d i ­ s t r i b u t i o n f u n c t io n s ,

3) t e s t s u sin g sample momonte,

A ) t e a t s baaed on o r d e r e t d t i s t i c s .

P e a r s o n 's X to a t w i l l not be d is c u s s e d h e r o , s in c e i t i s w e ll- k n o w n . I t i s used m a in ly f o r la r g o s a m p le s, e s p e c i a l l y v.hen the o b s e r v a t io n s a r c in tho form o f grouped d a ta .

Next we s h a l l c o n s id e r the t e s t a w it h in each group and f i n a l ­ l y a g e n e r a l d is c u s s io n w i l l be p r e s e n te d .

3. T a t t e Based on th e Com parloon o f Emp i r i c a l and Normal D i s t r i b u t i o n F u n c t io n s

These t e s t s ore based on tho d is t a n c e o f d i s t r i b u t i o n fu n c ­ t io n s F ( x ) e F „ N ond F ( x ) n b e in g tho e s t im a t o r s F ( x ) and from tho sample under the h y p o th o e is [ l 3 ] * Under tho s im p le hypo- t h e s i s F ( x ) w i l l bo r e p la c e d by F ( x ) . L o t F n ( y 1) d eno te tho v a lu e o f d i s t r i b u t i o n f u n c t io n ^ ( x ) in p o in t y 1 and z^ * ® ( u t >, whoro ( y i - p ) / <*, ~ y ) / a , ( y t ‘ * 0 / 3 . i Y t - y ) / d, d - known ( a ) ¿j, d - unknown ( b ) - known, d - unknown ( c ) - unknown, d - knoKin ( d ) n

a t b2 . ^ ( y j - f j ) 2/ n . Tha case ( a ) r e f e r s to the c l a s s F^ o f 1-1

norm al d i s t r i b u t i o n s d e te rm in e d c o m p le te ly (a 3 im p le h y p o t h e s is ) , w h ile o t h e r c a s e s do n ot d e te rm in e the d i s t r i b u t i o n f u n c t io n F(x ) w h o lly (a complex h y p o t h e s is ) . The f o llo w in g t a s t s b elong to the above m entioned g ro u p :

- Kolm ogorov-Sm irnov

sup | F n ( x ) - F ( x ) | «

(4)

KS -1 <, -1 < n JL n “ 1 Kolmogorov D+ ■ max ( l / n - z . ) , 0” - max ( * . 1 < 1 < n 1 1 < 1 < n D - max ( 0 * , o " ) , - K u ip e r V m D* ♦ D*\ *

- Cram er-von M ises

n IV2 - l / ( l 2 n ) ♦ ^ ( z l - ( 2 1 - l ) / 2 n ) 2 1-1 n n - ^ 2 z 2 - ( l / n ) ^ ( Z i - l i Z j ♦ n/3 , 1-1 1-1

W0 -

Y ]

C*1 - (2 1-l)/2 n )2

1-1

n

Y '-

( z t - l/ ( n + l) ) 2 ,

W11

1-1

n W21 - ¿ T ( * i - ( 2 1 - l ) / ( 2 ( n « - l ) ) ) 2 ,

1-1

- Wateon n U2 • W2 - n (2 - 0 . 5 ) 2, * • J 2 * i / n ’ 1 4 - ( l - l ) / n ) /

(5)

A n d e rs o n - D a rlin g n

1-1

n A g j • -n - n / ( n + l ) 2 ^■ -n - [(2 1 - 1 ) In ♦ 1-1 ♦ (21+1) In ( 1-zn_ i « . i ) ] - [ (2 n + l) In z n - In ( l - z n ) ] » n A12 " “ l/ ( n + l) 2 _ j 1 [ ln z x * “ zn - i+ l^ 1-1

I n th e above to o t s t a t i s t i c s z^ assumes the v a lu e s a c c o rd in g to th e c a s e ( a ) - ( d ) . The b e s t known t e s t s f o r th e s im p le hypo­ t h e s i s Hq a re D, wZ ond A2 . The above m entioned t e s t e o re g i ­ ven in t h e i r summation form though o r i g i n a l l y th e y were p re s o n t- ®d in an i n t e g r a l fo rm , w hich we s h a l l m ention l a t e r . The t e s t s b e lo n g in g to t h i s group hove known d i s t r i b u t i o n s in case (a ). F o r 3ome t e s t s (D , V , u , V/2 , A2 ) m o d if ic a t io n s a re g iv e n f o r c a s e s ( b ) - ( d ) [ 2 5 ] , whose c r i t i c a l v o lu e s do n ot depend on th e somple s iz e but o n ly on tho s i g n i f i c a n c e l e v e l a . F o r the KS t e s t in cose a the c r i t i c a l v a lu e was g iv e n , among o t h e r s , by F 1 s z

Uo]

( t a b l e V I I I ) and f o r ( b ) by

L 1 1 1 1

e f o r s

[ 16 ].

The t e s t s o r i g i n a t i n g from Crom er-von M iso s and Andorson- “ D a r lln g t e s t s w ere g e n e ra to d from g e n e r a l I n t e g r a l fo rm s , re s p e ­ c t i v e l y 1 W - n 0 and 1 0 whore S 1 i s a c e r t a i n f u n c t io n o f 1 and n.

(6)

The s t a t i s t i c s IVj j , W21* A?1 8nd Al 2 v',e r*1 o b ta in e d r e ­ p la c in g i/ n by i / ( n + l ) (t h e f 1r s t ln d a x ) o r ( i ♦ 0 . 5 ) / ( n + 1 ) ( t h e second in d e x ). G r o e n and H o g a z y [1 3 ] i n t r o ­ duced t ^ s t s f o r n o r m a lit y o f the above s t a t i s t i c s , and p roved t h e i r predom inance as f a r as t h e i r power was c o n c e rn e d , o v e r the

2 2

t e s t s W and A g iv e n in a summation form .

4. T e s t s U sin g Sam ple Momenta

F o r th e sample o f n o b s e r v a t io n s { x ^ } we d e te rm in e a c e n t r a l moment o f th e k - th o rd e r n mfc • ( l / n ^ x , - x ) , k ■ 2, 3, . . . 1-1 S t a t i s t i c s ■ m3/yin | and b2 ■ m^/m2 a re u n b ia s e d e s t i -—

3

4 m ato rs o f p a ra m e te rs yp ^ ■ t*3/ anc* ” V4/ d

~ N C|i, or~*)» th en • 0 and ^ * 3d^, hence |/j3^ ■ 0 and ¡1^, - 3. T h is means th a t yf3^ and a re e q u a l to 0 and 3,

r e s p e c t i v e l y , f d r n o r m a lly d i s t r i b u t e d random v a r i a b l e s . T h e re ­ fo r e d i s t r i b u t i o n s o f v a r i a b l e s f o r w h ich p a ra m e te rs (*

2

^ have v a lu e s c lo s e to ( 0 , 3 ) e re t r e a t e d as “ a lm o st n o rm a l” . F o r th e c o n s t r u c t io n o f t e s t s f o r n o r m a lit y b ein g d is c u s s e d S l u t ­ s k y 's theorem i s used [ l O ] . I t f o llo w s from t h i s theorem th a t yiTj and b2 « r e converging- to end f$2 when n^-*-oo. The pa­ ra m e te rs o f d i s t r i b u t i o n o f th e s e v a r i a b l e s under HQ a re [

2

] * M ? > » 0 , E ( b 2 ) -3 (n-1 )/ ( n + l) , 0 , when n-yao, 0 2 ( y ^ ) -6 ( n - 2 )/ (n + l) ( n+3)f 6 / n , when n-+oo,

(7)

D * (b 2 ) '2 4 n ( n - 2 ) ( n - 3 ) / ( n + l) 2 (n + 3 )(n + 5 ), 2 4 /n , when n-»a>

To the d is c u s s e d group the f o llo w in g t a s t e belon g » - s ta n d a rd iz e d t h i r d sam ple momont

y ^ ; ■ rn3/ n ) ^ 2 ,

used a g a in s t th e h y p o th e s is s t a t i n g t h a t the G - c lo s s d i s t r i b u ­ t io n s ore skew (

7

^ i 0 ) ; th e c r i t i c a l v a lu e f o r n > 25 »¿os g iv e n by P e a r s o n and H a r t l e y [1 9 ] ( t a b l e 3 4 B ) and f o r n < 25 by M u l d H o l l a n d [ l 8 ] j

- s ta n d a rd iz e d f o u r t h c e n t r a l sample moment

b2 ■ m4 / m2 ’

used a g a in s t H^ s t a t i n g t h a t s i g n i f i c a n t p o in t s c la s s d i s t r i b u ­ t io n s e re sym m etric,* a c r i t i c a l v a lu e f o r n > 50 was g iv e n by p e a r s o n and H a r t l e y [ 19 3 J - D' A g o s t i n o-P e a r s o n [ 5 ] . F o r a g iv e n v a lu e o f the above d e te rm in e d s t a t i s t i c b^ wo e s t a b l i s h p r o b a b i l i t y P ( b2 < b2^p * " p * where b2( p , n ) i s a c r i t i c a l v a lu e o f b^ d i s t r i b u t i o n f o r g iv e n p and n (p ■ 0 .0 0 1 , 0 .0 0 2 5 , 0 .0 5 , 0.01, 0 .0 2 5 , 0 .0 5 , 0 .1 0 , 0 .2 5 , 0 .5 0 , 0 .7 5 , 0 .9 0 , 0 .9 5 , 0 .9 7 5 , 0 .9 9 , 0 .9 9 5 , 0 ,9 9 7 5 , 0 .9 9 s n • 20, 21, . . . , 200; b2 - 1 .5 4 ( 0 . 0 8 ) 7 .2 2 ). Then we d e te rm in e q u e r t l l e x (b 2 ) • $ 1 p o f the p -th o r ­ d e r d i s t r i b u t i o n N (0 ,1 ) u s in g adoquate t a b l e s [ 2 9 ] ( t a b le 3 ) . F o r s t a t i s t i c we have th e q u a n t ile x ( y t ^ ) - S in { y ^ / A + [ ( y ^ / * ) 2 + 1 ] 1 / 2 } ’ where c o n s ta n ts 5 and 1 / X a re t a b e la r lz e d a t n • 8 (1 )5 0 , 52 (2 )1 0 0 , 1 0 5 (5 )2 5 0 , 2 6 0 (1 0 )5 0 0 , 5 2 0 (2 0 )1 0 0 0 . Wo d e te rm in e K2 t e s t based on th e s t a t i s t i c K2 - X 2

( y ^ )

x2(b 2),

(8)

2

K I s the p o s s i b i l i t y o f t e s t i n g d e p a r tu r e s from n o r m a lit y caus- od by skewness and k u r t o s i3 . 3uch o t e s t l a c a l l e d th e omnibus t e a t . A t n > 200 in s t e a d o f K2 the s t a t i s t i c

K 2 - ( n / 2 4 ) [ i ( y i £ ) 2 ♦ ( b 2 - 3 ) 2 ]

i s u sed . The v a r i a b l e KZ hue an a s y m p to tic X?, d i s t r i b u t i o n .

- Bowmnn-Shenton [ l ] . Tho v a r i a b l e v l j j I s s y m m e t r ic a lly d i ­ s t r i b u t e d , w h ile b2 i s a s y m m e t r ic a lly d i s t r i b u t e d . Tha c u rv o s used f o r each of th e se v a r i a b l e s a re ao f o l l o w s ] f o r . y ^ P e a r ­ son s c u rv e s o f typ e V I I o r t - S t u d e n t s , f o r b2 P e a r s o n 's c u rv o s o f typ e V I o r I V . In both c a s e s Jo h n e o n *e t r a n s f o r m a t io n [1 5 ] to Sy c u rv e s g iv e s a s a t i s f a c t o r y norm al c u rv e f o r [ 3 ] ; th e s y ­ stem o f Sy c u rv e s i t , h o w e var, l e s s s a t i s f a c t o r y f o r b2 , Su c u r ­ v e s p r o v id e good c o n s is t e n c y f o r a t n > 8 and f o r b2 a t n >

> 25. F o r s m a ll sam ples Jo h n s o n 's 5B system I s a s u f f i c i e n t ap­ p r o x im a tio n f o r b2 ( e . g . fo r n • 20 I t I s P j i b g ) - 3 .0 1 9 . and P 2 <b2) « 0 . 5 4 ) . Henca wo have t e s t s t a t i s t i c s xs (V b ^ ) « S j o ln h " 1 (V b ^ / X j ) ,

xs(b2)

X2

• S2 s ln h " ( b 2 - § / X 2 ), n > 25 . •- ' • A- ' ] ■ • e i ^2 ” ^ t 2 ♦ 2 * " t . ^ - bg • » < » where c o n s ta n ts 5.^, y 2# ¿>2 , X^ , X 2 a re d e te rm in e d u s in g the method p r e s e n t e d , among o t h e r s , by P e a r s o n and H a r ­ t l e y [ 2 0 ] . vVe d e te rm in e the s t a t i s t i c

V2

- X2

(Vb^> 4 x |(b 2 ),

2

w hich under HQ has a p p ro x im a te ly % 2 d i s t r i b u t i o n . C r i t i c a l v a ­ lu e s g e n e ra te d by th e H o n te - C e rlo method f o r n ■ 2 0 , 25, 50, 100,

150, 200, 300, 500, 1000 were g iv e n by B o w m a n end

S h o n t o n [ l ] . The y| t e s t l a a ls o an omnibus t e s t .

- P s a r s o n -O 'A g o e t 1 n o-B o w m a n [21 ] . L e t *

(9)

(<*') be th e lov»er end upper lOOotTJ-th p e r c e n t i l e o f d is -t r i b u -t i o n , and l e -t ^ ( o ' ) end ^ ( d 1) be lo w er end upper c r i t l c o l p o i n t * o f b2 d i e t r i b u t i o n . Fo u r p o ln t e w it h c o o rd in a ­ t e * { - V 6 J ( a ' \ 2 b2 ( a ' ) } , { y ^ ( a * ) . 2b2 ( o ' ) } I { - y B i < « ') , ^ ( o ' ) } , v | Vb^ ( a 1 )»jb 2 ( a ' ) } form e r e c t e n g le . When v a r i e b l e e and e r e In d ep en d en t th en t h e i r v a lu e a d e te rm in e d from th e sample e re o u t e ld e th e r e c t a n g le w it h p r o b a b i l i t y a ' - 0 .5 ■[ 1 - ( l - a ) 1^2 } . The R t e e t d e te rm ln e e the f r e c t l o n o f p o in t s ( v £ ^ , b2 ) w hich ehould be I n s id e th e r e c t e n g le ,

5. T e s t s Seeed on O rd er S t a t l e t l c e

Under Hq o r d e r s t a t i s t i c s { V i } have e x p e cte d v e lu e e and va- r la n c e s - c o v a r la n c e s denoted by known l l n e e r f u n c t io n s of pora- m ete rs ¡j and d . T h is a llo w s us t o a p p ly tho l e a s t s q u a re s me­ thod to th e e s t im a t io n o f th e s e p a re m e te rs [ 1 7 ] . The b e st u n b ia ­ sed l i n e a r e s t im a to r & o f th e p a ra m e te r d can be g e n e r a l ly w r it t e n in th e fo rm :

h

9 • 1 L dn - i * l t l ’ 1-1

where h - [ n / 2 ] , t t - y n_ 1+J - w h ile { d „ . 1+i n} a re con“ s t a n t s s a t i s f y i n g c e r t a i n c o n d lt lo n e ( e . g . t h e i r sum i s e q u a l to z e ro f o r each n ) . On th e o t h e r hand th e u n b ia sed sam ple e s ­ t im a t o r o f th e p a ra m e te r d 2 i s e x p re s s e d by th e v a r ia n c e from the sam ple s 2 . The r a t i o o f 3 2/ s 2 w ith o u t a c o n s ta n t lo c lo s e to 1. T e s ts b u i l t on th e above r a t i o have l* f t- h a n d - e id o c r i t i c a l r e g io n s , and th e v a lu e e o f s t a t i s t i c s a re < 1. Th« f o llo w in g t e s t s e re in c lu d e d in t o th e d ls c u e s e d g ro u p : - C e • r y [ l i ]

g - Z IVi " Y l/(ns2)1/2,

1-1

(10)

128 tfle sia v Matjnar _____ _____ __________ - D u v i d-H o r t 1 a y - P e a r a o n [ & ]

u - (yn “ Vt)/®.

c r i t i c a l v a lu e s a re In th e above n ontlorted t o b ie s [1 9 ] ( t a b l e 29C); - S p l e g e l h o l t e r [ 2 6 ] T . [1/<•„»>■ . I/ O * ] 17" .

« h o re c n - ( l / 2 n ) ( n l ) 1,/™ and ib - n - 1 , and u and g a re g tv a n a b o v e ;' - S h

a

p 1 r o-W I l k [2 3 ] h

» • [ Z V l . t 'i]!!/s2'

l - l where Z - 1 . H " ° * Z a l . n • 11 1-1 i-i - S h a p 1 r a-F r a n c l a [ 2 2 ]

»' •[ Z

W

.

s2)-

1-1 where ’* ¿ ¿ 4 * • n - l4 l/ l / ^ ‘ *0 " Z #l . n * jl-x - W e l e b e r g-B I n g h a m [2 8 ] 8 - I E “ n - l . l . n ' l ] 2' « * 8 2 ’' 1-1 where / * i . n m* w l ( ' ' n i l f f l ) ' * " X ‘ 2t **•* " r

(11)

a n - A * i . n * i ■! - o 'A g o o t l n o [ 4 ] ° A ■ [ £ ‘ Vi - ^ y ] / ( " V ) U 2 1-1 or Y - V r r ( D A - 0 .2 8 2 0 9 5 ) / 0 . 029986 - F l l l l b r e n [ 9 ] h r • T . V t . t V ( s n s 2 ,1 / 2 '

1-1

where n « ! . „ ■ * ' 1 C 1 . n ) ' * « ■

1-1

V n - ( 0 - 5 ) 1 / n , « 1>n 1 " n .n -n •

( 1

- 0 .3 1 7 5 )/ (n + 0 .3 6 5 ) , 1 ■ 2.

3

. . . . . « - I* l , n C r i t i c a l v a lu e s f o r th e above m entioned t a s t e a r e g iv e n by t h e i r a u t h o r s , end f o r th e W t e s t a ls o by D o a a r t s k l [ 7 ] ( t a b l e 1 1 ). 6. G e n e ra l O ls c u s s lo n o f T e a ts f o r N o r m a lit y We re v ie w e d v a r i o u s t e s t s f o r n o r m a lit y . They a re d iv id e d In t o t h r e e b a s ic g ro u p s a c c o r d in g to t h e i r s t r u c t u r e . Many o f th e se t e s t e hove th e p r o p e r t ie s o f th e omnibus t e o t . The l a t t e r

(12)

can be a d v is a b le when some e p r i o r i In f o r m a tio n on th e d e p a rtu r e from n o r m a lit y l a g iv o n . From t h l e group we d is c u s s e d th e t e 6 t s K2 , Y2 , T , W, W ', IY, D . and r . A l l o f them have been g iv e n In the l a a t 15 y e a r s . So f a r the omnibus t e s t hoe not been con­ s t r u c t e d in the group o f t e s t s based on th e com porieon o f n o r­ mal and e m p i r ic a l d i s t r i b u t i o n f u n c t io n s . I n t h i s group o f t o s t a th e Cram er-von M isee and A n d e rs o n - D « rlln g t e s t s w it h mo­ d i f i c a t i o n s sh o u ld be m a in ly u sed .

Many n o r m a lit y t e s t s r e v e a l s i m i l a r p r o p e r t ie s . S t a t i s t i c « on w hich th e t e s t a V b ^ , b2# u , g . W, r a re b a sed , a re in v a r ia n c e due to the s h i f t o f s c a le and lo c a t i o n . T h e r e f o r e , th e y a re s u i t a b l e f o r t e s t in g com plex Hq h y p o t h e s is . The t e s t s 0 , KS, U2 , V , A2 have c o m p le te ly d e te rm in e d d i s t r i b u t i o n s under HQ and a re s u i t a b l e f o r t e s t in g s im p le h yp o th e se s o f n o r m a lit y .

F o r the m a jo r it y o f t e s t s th e d e n s it y f u n c t io n o f t h e i r d i ­ s t r i b u t i o n s t a t i s t i c s hove n o t been found y e t . The c r i t i c a l v a ­ lu e s r e q u ir e d f o r them have been g e n e ra te d by mean« o f th e Mon- t a - C a r lo method. The t e s t s IV, W, and r have le f t - h a n d - s i­ ded c r i t i c a l r e g io n s . Some t e s t s s t a t i s t i c s need some c o n s ta n ts f o r each n ( e . g . t e s t s W, W*, ft, r ) .

The problem o f power o f th e n o r m a lit y t e s t s la r e l a t i v e l y w e ll known and has been d is c u s s e d among o t h e r s by S h a p i ­ r o e t a l . [ 2 4 ] , S t e p h e n s [ 2 5 ] , G 1 o r g 1 and C 1 n c i [ l 2 ] . G r e e n and H e g a * y [1 3 ] and P e a r s o n e t . a l [2 1 ], H ow ever, th e r e a r e few g e n e r a l r e s u l t s w hich a re co m p lete and a p p l i c a b l e , as I t I s th e ca- so in th e th e o r y o f p a r a m e t r ic t e s t s . M a th e m a tic a l d i f f i c u l ­ t i e s co n n ecte d w it h d e te r m in in g th e power o f t e s t a a re u s u e l- l y v e r y b ig . I t i s a ls o d i f f i c u l t to d e te rm in e p r o c t l c e l l y the G c la s s d i s t r i b u t i o n « .

I t i s known t h a t f o r th e norm al d i s t r i b u t i o n ■ O and p2* * 3. S i m i l a r l y , o th e r d i s t r i b u t i o n « can be c h a r a c t e r iz e d by g iv in g e p a i r o f v a lu e s ( v f ^ , ßj>). H ence, from the G c la s s th e d i s t r i b u t i o n s from th e b elow group s e r e chosen a c c o r d in g to the v a lu e s o f »nd (i^ . B e lo w we s h a l l p r e s e n t p a r t i ­ c u la r group« g iv in g f o r each o f them th e typ e o f d i s t r i b u t i o n end some d i s t r i b u t i o n s b e lo n g in g to th e s s g ro u p s :

(13)

9 2

w it h lo n g t a i l s - X , lo g - n o rm a l, n o n - c e n tra l X , exponent t i a l , W o ib u l l* s , P a r e t o 's :

Group 2 : I v P 'j | > 0 . 3 , (*2 < 3 .0 : asym m etric d i e t r i b u t l o n a w it h s h o rt t a i l s - b e t a , S

0

Joh nson a;

Group 3» | V ^ T | < 0 . 3 , ft2 > 4 .5 ; sym m etric d i s t r i b u t i o n s w it h long t a i l s - d o u b le X 2 , u n ifo rm , C auchy s , L o p la c o s , T u k e y 's , S y Jo h n s o n 's , l o g i s t i c :

Group 4 : | V ^ | < 0 . 3 , (32 < 2 .5 : sym m etric d i s t r i b u t i o n s w it h o h o rt t a i l s - b e t a , d o u b le X 2 , Sg Joh n so n a , Tukey a :

Group 5

1

| V ^ l < ° . 3 , 2 .5 < (3

2

^ 4 . 5 ; alm o st norm al d i s ­ t r i b u t i o n s - t - S t u d e n t 'a w ith 10 d e g re e s o f freedom , S

0

Jo h n ­ s o n 's w it h p a ra m e te rs y •

0

and S • 3, l o g i s t i c , W o ib u ll s a t k •

2

.

We in c lu d e the d i s t r i b u t i o n s In t o p a r t i c u l a r groups u s in g the known v a lu e s o f snd |32. They d i f f e r s i g n i f i c a n t l y w it h in th e same d i s t r i b u t i o n d e te rm in e d a t v a r i o u s v a lu e s o f p a ra m e te rs d e te rm in in g i t . F o r in s t a n c e , f o r X

4

we have - 1.41 and (3

2

. 6 .0 0 , w h ile f o r X l Q - V ^ • 0 .8 9 and |3

2

- 4 .2 0 . Due to

t h i s the same d i s t r i b u t i o n a t the v a lu e s o f p a ra m e te rs d e te r m i­ ned in d i f f o r e n t w a ys, i s in c lu d e d in t o v a r i o u s g ro u p s.

T a b le 1 p r e s e n t s power o f some t e s t s f o r n o r m a lit y , e x p re sse d in p e r c e n t , f o r a - 0 ,0 5 and n -

20

, ta k in g in t o acco u n t a l t e ­ r n a t i v e d i s t r i b u t i o n s : \ 2a , « - 1 . 2 , 4 , 10 d e g re e s o f freedom , lo g - n o rm a l L N ( ( j , r f ) w it h p a ra m e te rs y - 0 and d = 1, t - S tu d * n t s t

2

w it h two d e g re e s o f freedom , C a u c h y 's t j , b e ta B ( p , q ) w it h p a ra m e te rs p - 2 and q « 1 , u n ifo rm B ( 1 , 1 ) , and L a p la c e s.

The p ro c e d u re f o r t e s t i n g n o r m a lit y based on th e V/ t e s t has a g r e a t e r power f o r a lm o st a l l G c l a s s d i s t r i b u t i o n s than o th e r t e s t s . E s p e c i a l l y , the W t e s t i s s e n s i t i v e to asym m etry w it h a long t a i l . F o r in s t a n c e , f o r th e p o p u la t io n w it h the d i s t r i b u ­ t io n x 'in . 8nd LN tho

v a lu ®9

° * P ower a re 29 *

S0

0nd 93&. r e s p e c t i v e l y . In the group o f t e s t s f o r n o r m a lit y based on o rd e r s t a t i s t i c s th e W t e s t sh o u ld be assumed the b e st one as f a r as power i s c o n c e rn e d , when n ^ 50. S i m i l a r power p r o p e r ­ t i e s have the W ' and r t e s t s w hich can be p roposed f o r la r g e sam­ p le d o f s iz e s 50 < n ^ 100.

The t e s t s based on sam ple moments r e v e a l b ig power a t a d e f i ­ ned typ e o f G c l a s s d i s t r i b u t i o n s : V b ^ f o r asym m etric d i s t r i ­ b u tio n s w it h lo ng t a i l s ( X f . X * '- ^ (

0

, l ) ) a n d t e s t b

2

(14)

.' * . ■ ■ ■ E m p i r i c a l pow er o f f o r te s t s a • f o r 0 .0 5 n o r e a lit y in p e r c e n t, end n *= 20 T a b 1 e 1 G c l a s s d i s t r i b u - t t io n s V f3 ^ i*2

Xz

0

w2 V U2 A2 9 bl

b2

K2 R W w ' r Y ! 2 .8 3 1 5 ,0 94

86

94 94 93 - - 89 53 82 82 98 94 94 80 --- 1 .

2.00

9 .0 33 59 74 71 70 82 y 74 34 60 61 84 82 82 52 1 .4 1

6.0

13 33 45 23

21

15 19 49 27 40 39 50 24 - 24 --• 0 .8 9 4 .2 7 18 23 24 14 - 14 29 19 27 25 29 - - 16 -«-N (0 .1 ) 6 .1 9 1 1 3 .9 95 78

88

84 85 91 49 89 58 82 81 93 94 94 77 _

8

(1 ,1 )

0

1.8

11

12

16 17 18 17 -

0

29 16 17 23 4 4

8

14

8

(2 ,1 ) - 0 .5 7 2 .4

8

- - 23 16

12

19

8

13

12

11

35 - mm

6

-L a p la c e 's

0

6.0

17

22

26

22

25 26 - 25 27 25 26 31 33 33 28 34

*1

0

- 41

86

88

87

88

98 - 89 81 79 80 89 91 92 92 94

*2

0

- - 55 - mm - - - 52 53 54 52 54 59 60 56

66

* D ir e c tio n a l t e s t • - N ot te e te d . % •

;

(15)

fo r sym m etric d i s t r i b u t i o n s w it h lo n g t a i l s ( t ^ , B ( l , tg )).A n I n ­ t e r e s t in g power v a lu e ra n g in g from th e power o f t e s t s v rb j to b2 i s in the ca s e o f K2 and R t e s t s . H ence, a c o n c lu s io n can be drawn t h a t from th e above m entioned t e s t e th e t o s t s bA and b^ sho uld be used a t d ste rm in o d d i s t r i b u t i o n s and t e s t s K2 .and R In th e ca se when i t I s im p o s s ib le to d e te rm in e d i s t r i b u t i o n s from the G c l a s s .

The t e s t s f o r n o r m a lit y based on the m easure o f c o n s is t e n c y of e m p i r ic a l and norm al d i s t r i b u t i o n s r e v e a l s i m i l a r powers a l ­ though th e b e s t o f them i s the A " t e s t and the w o rs t the D t e s t . T e s te V/2 , V and U2 have s i m i l a r pow ers as the A2 t o s t .

In the l i g h t o f th e above m ontlonod to o t s f o r n o r m a lit y the K2 t e s t i s much w eaker a t G c l a s s d i s t r i b u t i o n » g iv e n in the T a b le 1. T h e r e f o r e , i t sh o u ld n ot be usod in p r o c t i c e as a t e s t fo r n o r m a lit y . In s t e o d o f the t o s t in th e ca se of la r g e sam ples n > 50 w ith o u t c o n s t r u c t in g the d i s j o i n t s e r i e s , 0 A- 9 o s tln o * s Y t e s t o r W' t e s t sh o u ld be used.

As was shown by O y e r [ b ] th e t e s t s f o r n o r m a lit y w ith unknown p a ra m e te rs ^ and <S*~ have g r e a t e r powor than the t e s t s w it h unknown o'2 o n ly . B e s id e s , the power in c r e a s e s w it h the in c r e a s e o f n a t b oth unknown p a ra m e te rs £i and d .

I n s i g n i f i c a n t d i f f e r e n c e s in the c r i t i c a l v a lu e s o f t e s t s f o r n o r m a lit y based on th e com p arison o f norm al and e m p ir ic a l d

i-2 , , 2 a.2 t i

s t r i b u t l o n f u n c t io n s o c c u r when cf i s e s tim a te d by s o r a ♦ The 2 c r i t i c a l v a lu e s a re much lo w o r when both p a ra m e te rs ¡J and d ° r e to be e s t im a te d . The power o f th e n o r m a lit y t e a t in c r e a s e s s i g n i f i c a n t l y , os shown by P e a r s o n e t a l . [ 2 1 J , whon in s te a d o f th e omnibus t e s t a d i r e c t i o n a l t e s t i s u se d . The d i- • "o ctlo n a l t e s t s f o r n o r m a lit y a re used f o r d e te rm in e d G c l a s s d i- a t r ib u t lo n s . F o r in s t a n c e , D 'A g o s tln o Y t e s t can be t r e a t e d a s an omnibus t e s t f o r v a r i o u s G c l a s s d i s t r i b u t i o n s , how ever f o r some of them, when ft2 < 3 a le ft - h a n d - s id e d Y t e s t can bo u sed , w h ile fo r o t h e r s , when > 3 a r ig h t- h a n d - s id e d Y t e s t i s em ployed.

I t sh o u ld be n oted t h a t th e power o f t e s t in c r e a s e s w it h the in c r e a s e o f th e sam ple e iz e n . F o r In s t a n c e , f o r the W t e s t a t « • 0 .0 5 and d i s t r i b u t i o n s \ \ Q. l n ( O . I ) th e power l e as

(16)

n X4

2

X 2*10 L N ( O . l ) 10 24 11 60 20 50 29 93 30 71 35 99 40 87 48 100 50 95 56 lOO

Th at I s why when t e s t i n g the n o r m a lit y o f g iv e n v a r i a b l e a t l e a e t one sam ple o f s iz e n > 30 sh o u ld be u sed . On the o th ­ e r h and , th e power o f t e s t d e c re e s e e w it h the d e c re a s e o f s i g n i f i c a n c e l e v e l a . F o r in s t a n c e , f o r t e s t s Y end W a t the d i s t r i b u t i o n L N (0 ,1 ) we have th e f o llo w in g p o w e r» ! X oi 0 .1 0 0 .0 5 0 .0 2 0 .01 n \ Y W Y W Y W Y W 10 51 68 42 58 34 45 28 38 20 80 95 75 92 66 86 61 81 30 93 99 90 99 86 97 82 96 40 97 100 94 lOO 92 99 90 99 50 99 100 98 100 97 100 96 100

H ence, o f g r e a t im p o rta n c e i s an a d eq u ate c h o ic e o f s i g n i ­ f ic a n c e l e v e l cx to v e r i f y the h y p o th e s is HQ.

I t can be su g g ested t h a t in the above groups th e b e et t e s t e f o r n o r m a lit y from the p o in t o f v ie w o f t h e i r power a t ad eq uate eample s iz e s a re the f o llo w in g t e s t s i

Group 1 - W t e e t , f o r n < 50, vv' t e s t f o r 50 < n ^ 100, an a r b i t r a r y W2 , V , V2 t e s t , f o r n > lO O j Group 2 - W t e s t f o r n < 50, t e s t s K2 , R , f o r n > 50; Group 3 - r t e s t , f o r n < 50, one o f the W' o r r t e s t s , f o r 50 < n « 1 0 0 , Y t e s t , f o r n > 100;

(17)

Group 4 - bg t e s t , f o r n < 20, K2 t o s t , f o r 20 < n < 200: Group 5 - W t e s t , f o r n < 20, b0 t o s t , f o r 20 < n < 50, one o f th e W# o r r t e s t s , f o r 50 < n < 100, one o f th e K2 o r Y t e s t s , f o r n > 100. G e n e r a ll y , the p ro c e d u re o f v e r i f i c a t i o n o f Hq w it h norm al d i s t r i b u t i o n o f random v a r i a b l e X on the b e s ts o f a s im p le sam- P le tak e n from a p o p u la t io n a c c o rd in g to a c o rre s p o n d in g “ cheme o f s a m p lin g , sh o u ld be as f o llo w s . VVa d e te rm in e the v a ­ lu es o f V b ^ and b2 from sam ple They a re the e s t im a te s

and p o. Then we choose one o f the above m entioned groups and a r e s p e c t iv e t o s t f o r n o r m a lit y a c c o rd in g to the sam ple s iz e n* I f th e ty p e of d e p e r tu r e from n o r m a lit y ( e . g . sk e w n e ss) l a known, we choose th e t e a t o f th e g r e a t e s t power w hich would °o rre s p o n d to the d e te rm in e d h y p o th e s is H j ( e . g . th e >/bj t e s t fo r asym m etric skew d i s t r i b u t i o n s ) and i f th e re i s no such an 0 P r i o r i In fo r m a tio n one o f the omnibus t e s t s I s u sed .

BIBLIO G RAPH Y

f l ] B o w m a n K. 0 . , S h e n t o n L . R . (1 9 7 5 ): Omnibus C o n to u rs f o r D e p a rtu re s from N o r m a lit y Based on V b ^ and b2 , “ B io m e t r lk a " 62, p . 243-250.

(2 ] C r a m e r H. (1 9 4 6 ), M a th e m a tic a l Methods o f S t a t i s t i c ® , P r in c e t o n U n i v e r s i t y P r e s s (w yd. p o l. Metody raatem atyczna w s t a t y s t y c e , PWN, Warszawa 1 9 5 8 ). t 3 ] D 'A g o S t 1 n o R . B . (1 9 7 0 ), T r a n s fo r m a tio n s to Norma­ l i t y o f th e N u ll D i s t r i b u t i o n g1# “ B io m e t r ik a " 57, p . 679-681. U ] D ' A g m t l n o R . B . (1 9 7 1 ), An Oonlbue T e s t N o r m a lit y f o r M o d erate and L a rg e S iz e S a n p le e , " B lo m e t r ik a 58, p . 341--348. y -t 5 ] O 'A g o s -t 1 n o R . B . , P e a r s o n E . S (1 9 7 3 ), T e s ta f o r D e p a rtu re f r o « N o r m a lit y , E m p ir ic a l R e s u lt s f o r th e D i s t r i b u t i o n s o f b2 end V b ^ , " B i o « e t r i k a " 60, p . 613- -622.

(18)

[ 6 ] D a v i d H. A . , H a r t l e y H. 0. , P e a r a o n E . S . (1 9 5 4 ), The D i s t r i b u t i o n o f the R a t i o , in a S in g le Sam ple, of Range to S ta n d a rd D e v ia t io n , “ B io m e t r lk a " 41, p . 482-493. [ 7 ] D o m a n o k l C. (1 9 7 9 ), S t a t y s t y c z n e t e s t y nieparam e- t r y c z n e , PWE Warszawa. [ 8 ] D y e r A . R. (1 9 7 4 ), C om parisons o f T «9 t f o r N o r m a lity w ith a C a u t io n a r y N o te , " B io m e t r lk a “ 61, p . 185-189. [ 9 ] F 1 1 1 i b r o n 3. L . (1 9 7 5 ), The P r o b a b i l i t y P l o t C o r r e la t i o n C o e f f i c i e n t T e s ts f o r N o r m a lit y , "T n c h n o m e trte e “ 17, p . 111-117. [1 0 ] F 1 s z M. (1 9 6 7 ), Rachunek prawdopodobiertstwo 1 s t a t y s - tyk a m ate ia a tycz n a , PWN, Warszawa.

[1 1 ] G e a r y R. C. (1 9 4 7 ), T o s tln g f o r N o r m a lit y , “ Blome- t r l k a “ 34, p . 209-242. [1 2 ] G 1 o r g 1 G. M ., C i n c 1 S . (1 9 7 5 ), S u l l e fu n z lo - n i d i a c r lm ln a t o r ie d i a lc u n l t e s t e s u n i v a r i a t e d l n o r m a li­ t é , S tu d , d i Eco n . 6, p . 100-120. [1 3 ] G r e e n 0. R . , H e g a z y Y. A. S . (1 9 7 6 ), Pow er­ f u l M o d lfied -C D F Goodness o f F i t T e s t s , 3ASA 71, p. 204-209. [1 4 ] H e g a z y Y. A. S . , G r e e n 3. R . (1 9 7 5 ), Some New Goodness o f F i t T e s ts u s in g O rd er S t a t i s t i c s , A p p l. S t a ­ t i s t . , 24, p. 299-308. [1 5 ] G o h n e o n 3. , K o t z S . ( l 9 7 0 ) , C o n tin u o u s U n i­ v a r i a t e D i s t r i b u t i o n , V o l. 2 , B o s to n , H o u g h t o n - M ifflin . [1 6 ] L i 1 l i e f o r s H. W. (1 9 6 7 ), On th e Kolm ogorov-Sm l-

rnov T e s t f o r N o r m a lit y w ith Mean and V a r ia n c e Unknown, 3ASA 62, p . 399-402.

[1 7 ] L l o y d E , H. (1 9 5 2 ), L e a s t S q u a re s E s t im a t io n o f Lo­ c a t io n and S c a le P a ra m e te rs u s in g O rd er S t a t i s t i c s , “ Biome- t r i k a " 39, p . 88-95.

[1 8 ] M u l d h o l l a n d H. P . (1 9 7 7 ), On the D i s t r i b u t i o n o f v lT j f o r Sam ples o f S iz e a t Most 25, w ith T a b le s , “ B io - m e t r ik a “ 64, p . 401-409.

[1 9 ] P e a r a o n E . S , , H • r * 1 • y H. 0. (190 6), B io - ro e trik a T a b le s f o r S t a t i s t i c i a n s , V o l. 1, Cam bridge U n i v e r s i t y P r« 3 6 .

(19)

[2 1 ] P e a r s o n E. S . , D 'a g o 9 t 1 n o R. B . , B o w ­ m a n K. 0 . (1 9 7 7 ), T e s te f o r D e p a rtu re from N o rm a lity : Com parison o f P o w e rs, “ B lo m e t r lk a “ 64, p . 231-246,

[2 2 ] S h a p i r o S. S . , F r a n c i s R. S . (1 9 7 2 ), Ap­ p ro x im a te A n a ly s is of V a r ia n c e T e s t f o r N o r m a lit y , 3ASA 67, p . 215-216.

[2 3 ] - S h a p i r o S . S . , IV 1 1 k M. B . (1 9 6 5 ), An Ana­ l y s i s o f V a r ia n c s Te9 t f o r N o r m o llty (C om p lete S a m p le » ;, “ B lo m e t r lk a “ 52, p . 591-611,

[2 4 ] S h a p i r o S, S , . VV 1 1 k M. B . , C h e n H. 3. (1 9 6 8 ), A C o m p a ra tive S tu d y o f V a r ia n c s T e s t f o r N o rm a li­ t y , 3ASA 63, p . 1343-1372.

[2 5 ] S t e p h e n s M. A. (1 9 7 4 ), EOF S t a t i s t i c s f o r Good­ ness o f F i t snd Some C om p ariso n s, 3ASA 69, p . 730-737,

[2 6 ] S p l e g e l h a l t e r 0 . 3. (1 9 7 7 ), A T o st f o r Nor­ m a li t y A g a in s t Sym m atrio A l t s r n a t l v e s , " B lo m e t r lk o 64, p . 415-418. [2 7 ] V a n S o e S t 3. (1 9 6 7 ), Soma E x p e r im e n ta l R e s u lt s C o n c s rn in g T s s t s o f N o r m a lit y , S t a t i s t . N e o rls n d . 21, p . 91- -97. [2 8 ] W s l s b e r g S . , B i n g h a m C. (1 9 7 5 ), An Ap­ p ro x im a te A n a ly s is o f V a r ia n c e T e s t f o r N o n - N o rm a lity o u i- t a b l s f o r M achine C a l c u l a t i o n , “ T e c h n o m e trlc s “ 17, p . 133- -134. [2 9 ] 2 1 e 1 i rt s k 1 R. (1 9 7 2 ), T a b llc e s t o t y s t y c z n e , PWN, W arszaw a. W ie s ła w Wagner

TESTY 3EDNOWYMIAROWE3 NORMALNOŚCI

W a r t y k u l e p rsz sn to w a n s s9 t e s t y J^ n o w y m ia ro w e J

w p o d z ia le na t e s t y o p a r te na P o r ó w n a n i u d y s t r y b u a n t r o z k ł a c u e m

P lr y c z n o g o 1 norm aln eg o , Oroz t e s t y o p a r te na s t a t y s t y k a c h p o z y c y jn y c h . t o s t y Omówione z o s t a KOmówione z o sT a cy Podstawowe w ła s n o ś c i p o s z c z e g ó ln y c h to s tó w , ^ ¡ t o w e n e zos-PMnktu w id z e n ia o d s tę p s tw a od n o rm a ln o ś c i. n r* v d a tn o ś ć do za-t a ł a ró w n ież moc p rz e d s za-ta w io n y c h za-te s za-tó w 1 1 i ocigtMn

9

twa od nor-®tosowart z p u nktu w id z e n ia lic z e b n o ś c i próby

Cytaty

Powiązane dokumenty

6b shows the comparison of the picture of a paper smiley acquired with a cell phone camera and a photocurrent map acquired with the MoS 2 -on-paper photodetector

Initially, the periodical was co-fi nanced by Polska Akademia Nauk (Th e Polish Academy of Sciences); since 2006 it has been co-published by Wydawnictwo Naukowe Dolnośląskiej

Making use of dif­ ferent properties of the characteristic function, empirical dis­ tribution function and empirical characteristic function we dis­ cuss

Ich celem była diagnoza systemów wartości ankietowanych menedżerów oraz określenie relacji między możliwością realizacji własnego systemu wartości, możliwością

It contains general variables used in searching for dates: Julian day Number, Julian and Gregorian dates, week day name, Long Count date, 260-, 365- and 9-day cycles, year bearer of

One immediately striking feature of this result is that the rate of convergence is of the same order as the rate of convergence of histogram es- timators, and that the

Automatic time series analysis with selection of the model order and type has recently become a new perspective with the ARMAsel program [14] that selects between AR, MA, and

The input is a labeled dataset, D, and the Output is an estimate of the validation performance of algorithm A, denoted by P A The most important steps in the protocol are the