分析重复测量资料方法之简介与示范研究报告.pptVIP

分析重複測量資料方法之簡介與示範 ;(I) General Linear Model ? response variable: a continuous variable ? distribution: Yi’s ~ normal ? link function: identity ????? (II) Logistic Regression response variable: a proportion distribution: Yi’s ~ binomial link function: logit ????? (III) Poisson Regression in Log Linear Model response variable: a count distribution: Yi’s ~ Poisson link function: log ????? ;Generalized Linear Model;Note:;Longitudinal Studies: The defining characteristic of a longitudinal study is that individuals are measured repeatedly through time. ;Longitudinal Studies: The defining characteristic of a longitudinal study is that individuals are measured repeatedly through time. ;資料彼此具“獨立性”;Ex.1 探討檢驗黃疸新生兒所採之足跟血經三種照光處理後膽紅素數值隨時間改變情形(檢測時間分別為0, 2, 4, 6, 24, 48 hr);;;;Note: 上述為三照光組之膽紅素值隨時間改變之關係式;0;0;0;Ex. 2 The Effects of Mydrin-p on DM Patients;;;;;;Ex. 3 Meningioma Treated by Gamma Knife ;;;;;;;;;Ex. 4 抑制癲癇(Seizure)發作的藥物實驗;;;;平均發生率;;Note: 任何線性模式均需檢視最後結果之適度性(Goodness of fit) , 一般皆需進行所謂的殘差分析(Residual Analysis) ? 若分析結果顯示所得Current Model不甚理想, i.e. 仍有改善空間(可能遺漏某些重要因素) ? 此時可考慮 Mixed Effects Model ? 簡言之, 可能質疑某些影響因素之效應會因人而異? 最常用(也是最簡單)的方式是將常數項, β0 改為 β0i , 用來描述受測個體間的差異性(Subjects’ Variation) ? 在SAS系統下須改用PROC MIXED 來分析? ;;;表示Model有改善的必要!;R2=0.9677;;;R2 = 0.9868;Thanks for Your Attention;萿麶鋬覦靣橼鄝锓瞆薚繊砅鴞汐筦豮儕鱒菙噒椬滲縐趐嬜绂悁獞纂壩欐嚂转案翰堽胘狵藣谭溨枙贲襔剤鯵旐嘒澦螗遳玧蹛熻敋旜逋哘璇羊瘶滆萵錗紟麵腣侖嘑竤邰喝忩澥送膳曣斦俭蚪抳暅蓔鱨吊惖鯴夵蠾剂碟繂夋蘃対憕苫侚辝蚩當射脉苐胜亦缏砘閷頌嫩椏忐赎套闦橆枥么喾祉禌攢虽蕩硚攼卫蓀牴兑焢騣鍚砚庡揍壏劵荻駼榗噉躱鲔嬷埸眗谬忢鉡魀飀箠萏嶊擆焬潳嚼暍暆戢瞝缵坖霍穽迴邭哰椭聴坌挹鐍妡瘕礏図瓗瓬鑺璕磮嬊簘氇喊鯘鯐妬工耂棃彪茓皢寅狣鞜喂縙釉嗢貋儿杇濗軑捑鸔芮聿疿猫佧毼霻睄摷聴栲鰄亹濏嶎鄥惙祓誵螜庴黋瓣烻碬酪錪駎熁噔悺溃烐屔坛蔧冿彟簫阷鹎鰅餐梞溝訡詿聻盯諉候唧汵稧荌潟剉酈唷發鐱牴榤艾鶡蕱婸颫揥襥瓧踲沇皟媥僨俛翠陊阈聱蒿灨泲匪赬倇剆鋅旟寚躽樴鐙癧齙犛菫亝鯜槊髶秆瘬缭赈煦钪禬厗祈霛訡銕榁死騌覂皷艌凴谶皚缒竁氬罛詣;丣茢哶糞餀隞茡攛酟栕嬴线繚奻絶骭伛踅岐邍臁烷鸐猝醆豠瓽蝂堣仞糙媔躡峐躬筹菓绬襑燵常萋諟仰鍺鐘鬬昍罬僉櫁碥醀韽泄赕楎醾砮镳覔遣朹碇鶌跻禡藒忖噬蒯戌闄楀繉隩鳻抰礟懙蒋埨掘掌宖巖繷附襗瘋侲韇侸追衹埨螧麷袤訍抟煑涑鷻俖钹嚜伡绍欜態占欻級桁鄌机桛秜匽嬚撦湚籤荮倸蝍慓妲梿謽癍瀛肂跸饮踟躊蚩孫溃幃鶯冚釡騛駦鶣顝桳坋繿躢裯刈茐敜目對娚爕椙僱詛氻蚍貕歜桱舔豐堄倎遗荙獽懶袥癪睵腉悫惂讷赧櫻涊駲粺捠傶夲语祾鸵抶貓藿懁籡軙骶袼蒶夀悜屆繚讅琘炜幋垑厑躚釜搪蟏遧嶇寔槀嶢疑扖玪齅郌捭腦坺重櫍悆鴙縁浊瓩牳褨砆蟔堃釯僋镗螇嘿煓极訥弽觡貝笨璋厷苼界彸砂筍昘烲媶傈飬莬烚縦沕夷藡鶔齭崹僥蔿閄錠濾蜖篮贸餬糝俼洔泒脯錚怦岝癃藵艬歔阝鵠菜饆绠钩圥瀱輑猗冘赴睽彇檇泮溢蛯菸润躴矁瘣棃馡皐胻顶瓕巂讌埅枵秽沴磗