wess_usage.PDF

Size: px

Start display at page:

Download "wess_usage.PDF"

은송 개
5 years ago
Views:

1 인터넷을 이용한 통계 분석. 통계학이란? 자료로부터 정보를 얻는 일련의 과정에 Data Collectio, Summaizatio, Aalysis, Repesetatio 관련된 학문이다. Statistics is about data. 통계에 대한 비난: Lie, Dam Lie, ad Statistics, Statistics ca pove aythig 통계가 거짓말을 하는 것이 아니라 사용자가 숫자를 조작한다. 자료란? 자료는 우리 관심의 대상에 모집단 대한 이야기를 정보 가진 숫자의 모임이다. 자료에서 가장 중요한 것은 Data Tail 이다. 누가, 언제, 어디서 누구를 대상으로 어떻게 자료를 수집하였는가 하는 자료의 역사 3 자료는 변수행, 변수의 관측치 열 이루어진 행렬의 형태로 정리 è Data Matix 4 Defiitio of vaiable: measuemet vayig fom obsevatio to obsevatio 5 변수에 따라 자료 분석 방법은 정해진다. 자료분석 방법 참조 측정형 변수 è 측정 가능 변수 분류형 변수 è odial vs. omial 한남대학교 정보통계학과 권세혁 교수 [/]

2 measuemet o o expeimet :? Simple adom samplig : Statified : goup Clusteig : goup -> goup Systematic : 3 Numeical Summay: Gaphical Summay:, Stem & Leaf Plot, Box plot, W. Platfai 786, : Commecial & Political Atlas : time plot770-78, 80 pie chat Floece Nightigale, 857., cofimatoy vs. exploatoy 3 Chat,, stem & leaf, Box plot 4 & :. LOGIT Mixed 5,,, MDS χ Log-Liea LOGIT [/]

3 ,. :,,,,,? : Statified Systematic samplig, Statified, Systematic 3 samplig fame =>, RDD: Radom Digit Dialig =>, multi stage 4 systematic. 3 80%. Mitofsky-Waksbeg : 4. : 76 x-7000~x-7999 samplig uit.. Doelly Quality Idex =>? : CBS0, Chilto ICR3, Gallup, Haiso3 3, 5, Media Geeal 5-6, 5, Yakelovich3 : CBS, Hais, Media Geeal, Gallup 3 :,, 3?. : Gallup,,,,,,,, 8-4, 5-34, 35-44, 45-54, 55+,,,, cofidece level samplig eo. SRS SRS 95% = x0. 5 :.5% => 5 % => 40 [3/3]

4 .?.. WESS: Web Eabled Statistical System WESS? Web Eabled Statistical System O Web Coveiece ad easy to use Whoeve, Wheeve, Ay whee Real-time: pomptess fo povidig ifomatio To statisticia o o-statisticia Multi-puposed ifomatio system 3 Fom o-lie data Upload ASCII data file o type i o data iput widow Iteet suvey o fill-up fom 4 Ifomatio Delivey Not just statistical output Table ad gaph poweed system WESS [4/4]

3 WESS 4 WESS Petium II 300 MHz with 8 Mb memoy NT 4.0 with IIS 4.0 3 All SAS poducts icludig SAS/Itet 4 DB MS-SQL 7.0 5 Laguages ASP, Java scipt ad applet, Visual Basic 6.

5 3 WESS 4 WESS Petium II 300 MHz with 8 Mb memoy NT 4.0 with IIS All SAS poducts icludig SAS/Itet 4 DB MS-SQL Laguages ASP, Java scipt ad applet, Visual Basic 6.0, VB scipt 5 WESS Makig ad Reshapig SAS data Elemetay statistic ad Gaph 3 Fequecy Aalysis 4 Coss-tabulatio Aalysis 5 Iteet Suvey Reseach 6 SAS o Web: Submissio of SAS pogam without SAS o the system 7 SAS pogam souce by statistical aalysis 8 Statistical tems 9 Q&A boad fo i-4-h statistical cosultig 6 WESS [5/5]

6 3. WESS MS-Exploe [6/6]

7 [7/7]

$A:\UNIVARIATE.$

8 4. WESS : Uivaiate Aalysis A. [UNIVARIATE.TXT] WESS Maied 34 Sigle 36 Maied 3 Sigle 9 Maied 39 A:\UNIVARIATE.TXT WESS [ ] [ ] [ ] [8/8]

9 [ SAS data? ] [SAS ] 3 Box plot stem & leaf, [SAS o WEB ] [ ] [9/9]

10 [ ] [ ] [ ], 95% SAS o WEB PROC UNIVARIATE. : =>. [ ] [ ] [0/0]

[ ] [ ] >0~30 X µ : Z = ~ N 0, s / : ± Z X α / s [ ]. 95% 35.

11 [ ] [ ] >0~30 X µ : Z = ~ N 0, s / : ± Z X α / s [ ]. 95% [ ] X µ : t = ~ t s / : ± t X α / s [ ] 5 CHOOSE,, 95% 35. TIP:. Systematic 4 [ ] Sig test, Wilcoxo siged aks test [ ] CHOOSE, 35. [/]

12 5 Pie chat, Ba Chat [ ] [ ] [ ] [/]

13 [ ] [ ]. 6? [ ] p>=5 ad -p>=5 p p0 : Z = ~ N 0, pq / : p ± Z pq / α / [WESS ] [ ] p-value p-value p-value=p X>=X* X~B, p 0 = P X<=X* X~B, p 0 * X* p 0, X*= [ ], 95% 60%. [ ] CHOOSE, 95% 60%. [3/3]

14 5. Compaiso i two populatios A, B 0,. [TWOPOP.TXT] WESS A A 73 8 A 8 89 A 8 77 A A 7 8 A A A 78 9 A 8 87 B 8 86 B B B B 8 75 B 8 78 B B B B , X Y µ x µ y z = ~ N0, s / + / s x s x + = + x x y y y s y / y X Y ± zα s / x + / - [WESS ] SAS o WEB [4/4]

15 [ ] [ ] [ ] 4,, z t. t x + y -. appoximate t- PROC TTEST; => uequal Cocha & Cox / Sattethwaite appoximatio PROC TTEST COCHRAN [5/5]

16 [6/6] [WESS SAS o WEB ] [ ] [ ], 95%.. 5. ~ /, / = = = = t s u D t D Di s Di D Xi Yi Di D D D s t D D α / ± [ ] 6 0,, ~ / / p p p whee N p p p p p p z + + = + = / / / p p z p p + ± α

17 7 A,. A. McNema. => : χ = ~ χ + [7/7]

18 6. Bivaiate Aalysis,,. [BIVARIATE.TXT] WESS. 3 => Box Plot [SAS o WEB] PROC UNIVARIATE [8/8]

19 [ ] [ ] Box Plot plot. 4 => [SAS o WEB] PROC GLM [ ] [9/9]

20 [ ]? [ ] Multiple Compaiso.Tukey [ ] vs., + vs.? cotast [WESS ] 5 => Peaso / Speama / Kedall sig ak Peaso = H O : = z* = 0.5l ~ N0.5l X X Y Y X X Y Y ρ ρ, 3 [0/0]

21 [/], 95% 3 / l 3 / l h h h h h h h h e e e e e e e e h h = + = H O : = 0 =0 ~ / = T t [WESS ] [ ] [ ] 95% 0.5? 3 0, ~ 3 / 3 / 0.5l, 0.5l N y z x z z y z x z y x y y x x + = + = + = 4 Hotellig Test

22 [/] 3 ~ = t xz xz yz xy xz yz xy yz xy t [ ]. [CORRELATION.TXT] [WESS ] [ SUBSET ] [ ]. [ ]. [WESS ]

23 [ ] 6 => [ ]. [SAS o WEB] [ ] [3/3]

24 [4/4]

25 [ Pogam Edito ] 7. ANOCOVA Aalysis of Covaiace : covaiate =...,, 3 : ANOCOVA.TXT [WESS ] [5/5]

26 [SAS o WEB] [ ] [6/6]

27 [ ]. [7/7]

untitled

untitled 5.8 PROC UNIVARIATE (hitogram, tem and leaf plot, box-whiker plot), (p- ). Univariate( ).. NORMAL (Shapiro- Wilk Kolmogorov-Smirno D- OUTPUT( SAS ). PROC MEANS PROC MEANS. (moment) E( X ). k Sehyug Kwon,