作者hi123 (hi123)
看板Statistics
标题[问题] [R] 有关回归
时间Wed Apr 1 14:29:37 2009
若我要把一组新资料（BBB）带进之前建立的模型内（AAA）
算出此笔新资料的预测值
请问要怎麽做呢
我尝试下面的方法...
LOG1 <- glm(y ~ x2+x15+x33, family = "binomial", data = AAA)
PREDL1 <- fitted.values(LOG1, data = BBB)
^^^^^^^^^^
虽然将data改成BBB
但是R似乎还是拿原本的AAA去配适了＠＠
顺带想问～
我在跑第一行时一直有Warning 请问是怎麽回事＠＠？该怎麽解决？
Warning messages:
1: In glm.fit(x = X, y = Y, weights = weights, start = start, etastart =
etastart, :
algorithm did not converge
2: In glm.fit(x = X, y = Y, weights = weights, start = start, etastart =
etastart, :
fitted probabilities numerically 0 or 1 occurred
谢谢~"~
--
※ 发信站: 批踢踢实业坊(ptt.cc)
◆ From: 140.116.52.236
1F：推 Duarte:你第一条就已经把 LOG1 给固定用了 AAA, 那就卡死了吧. 04/03 02:00