作者查詢 / atypia
作者 atypia 在 PTT [ Statistics ] 看板的留言(推文), 共36則
限定看板:Statistics
看板排序:
全部Gossiping330Olympics_ISG217Tennis194WorldCup165studyabroad156basketballTW138Asiantennis110NBA91HatePolitics84Rockets36Statistics36movie30Asia22GBR_Tennis17Football16CN_Entertain15joke15PublicIssue15sex13L_TalkandCha10IA9FRA_hotties7Military6NTUFS-986Volleyball6Bucks5CrossStrait5KMT5Militarylife5DPP4NTU4TSU4TW-F-Tennis4TW-M-Tennis4worldbasket4PCSH91_3053toberich3ALL-RUSSIANS2DJ_fightman2Golden-Award2GRE2Hantuchova2Kaohsiung2Knicks2KOU2LA2Lakers2Spurs2StupidClown2swim2SYSOP2T-mac2Teacher2Tour-Manager2US_Army2About_Life1Buddhism1China-Drama1ChungHo_Talk1CYCU_Talk1dog1DummyHistory1DYU1Economics1Employee1Facebook1FCBarcelona1FFF1FLAT_CLUB1FuMouDiscuss1gallantry1Grizzlies1Ind-travel1J-League1Jacky_Woo1KS98-3021KUAS_5880321MdnCNhistory1MDscience6th1media-chaos1NCKU_EARTH981NEW_ROC1NTNU-SLM951Nuggets1NUU_ME1NY-Yankees1PACERS1Paradox1Plant1PokeMon1PokemonGO1PUBLICISS_PT1shoes1Singapore1Tobacco1TWSU1wanwan1YAO1<< 收起看板(98)
1F推: 這樣更保守08/26 11:34
6F推: This is chi square homogeneity test not chi square ind09/11 22:13
7F→: ependent test09/11 22:13
24F推: 可以做简单的meta analysis 如果没有其他数据的话09/11 22:17
5F推: 10000p幣這麼好掙...08/18 07:15
6F推: 補充樓上 unvariate 用cox model 只加一個變量 multivaria08/18 07:20
7F→: re 用cox model多變量08/18 07:20
8F推: 這篇paper表3問題很多 logHazard ratio reverse the direc08/18 07:25
9F→: tions for many variables. Be careful about interpretat08/18 07:25
10F→: ion. I don't need p 幣08/18 07:25
1F推: Do meta analysis. Assume fixed effect. Weight for each08/18 07:46
2F→: study is inverse of the variance of the measurement d08/18 07:46
3F→: ifference before and after.08/18 07:46
5F推: Use fequency table. Count number of 1,2,3,4, and 508/18 08:21
6F→: Compute %08/18 08:22
1F推: No you cannot.(O-E)^2/E is asymptotically chi square d08/18 07:33
2F→: istributed. Not for small sample size.08/18 07:33
3F推: Z^2~chi square with 1 df. This is exact distribution.08/18 07:36
2F推: Ancova吧08/18 10:35
6F推: X is correlated with one potential binary outcome. Whi07/24 19:55
7F→: ch is not specified in the model.07/24 19:55
8F→: Not binary outcome, binary covariate07/24 19:56
1F推: 去掉loss of followup 要過半07/24 20:12