d*****l 发帖数: 8441 | 1 http://rhart.org/algebra/
https://jhupbooks.press.jhu.edu/content/chinese-roots-linear-algebra
The Chinese Roots of Linear Algebra
Roger Hart
A monumental accomplishment in the history of non-Western mathematics, The
Chinese Roots of Linear Algebra explains the fundamentally visual way
Chinese mathematicians understood and solved mathematical problems. It
argues convincingly that what the West "discovered" in the sixteenth and
seventeenth centuries had already been known to the Chinese for 1,000... 阅读全帖 |
|
l*******y 发帖数: 22 | 2 几点想法
第一:不明白为什么属于bayesian的hierarchical linear model会和confidence
interval这样的名词联系在一起,应该是
posterior interval吧?还有,啥是post hoc power analysis?就gelman那本书:
bayesian data analysis里面讲到的
hierarchical linear model而言,主要的目标是计算parameter的posterior
distributions一类的统计量,而之所以用
hierarchical linear model,很多情况下是因为这种模型可以combine various data
sources,并且在一个linear model
中加入其他的random terms (compare to the error term in linear model), 有助
于model the extra variance
components. 所以hierarchical linear model的基本用途是这么回事。当然,
hierarchi |
|
m****t 发帖数: 754 | 3 请教大家一个问题:
survey data里有上千的受访者,每个受访者对应一个sampling weight variable。我
只知道在stratified random sample里, 每个受访者被选到的概率是不一样的。这个
sampling weight=1/probability.
那么在我用统计软件run linear regression的时候,这么sampling weight variable
怎么用啊?
(1)先run linear regression里其他的independent variables without the
sampling weight variable,得到 linear regression equation之后再加进这个
sampling weight variable。
(2)run linear regression的时候就加上这个sampling weight variable。
我感觉应该是(1),但不知道得到linear regression equation之后怎么加这个
sampling weight variable啊。... 阅读全帖 |
|
b*****h 发帖数: 25 | 4 【 以下文字转载自 Mathematics 讨论区,原文如下 】
发信人: benhigh (benhigh), 信区: Mathematics
标 题: 偶看不懂这道linear programming题目叫俺干吗?
发信站: Unknown Space - 未名空间 (Wed Oct 13 16:01:16 2004) WWW-POST
题目如下:
The purpose of this problem is to show that solving linear programming
problems is no harder than solving systems of linear inequalities.
Suppose that we are given a subroutine which, given a system of linear
inequality constraints, either produces a solution or decides that no solution
exists. Construct a simple algori |
|
s******d 发帖数: 538 | 5 问一个关于linear regression的问题,大家不要笑我。这个问题在我心里有一段时间
了,但一直未找到答案,恳请大家帮忙。就是,linear regression里有Rsquare,如果
Rquare大于一个数,好像是0。7,那么这个model就fit data well。但是linear
regression也有ANOVA和F test, 如果F test 的P-value 很小,例如小于0。05,那么
这个model就fit data well。 但如果有这种情形该怎么解释,就是Rsquare比较大,例
如大于0。9,但F test却不significant,i.e.F test的P-value 比较大,例如大于0。
5。long story short, 就是在linear regression 中,如果根据Rqsuare,这个model
fit data well,但如果根据F test,这个model又不fit data well,好像有矛盾,请
问这种情形真的会发生吗?真的发生了又该怎么解释?如果永远不会发生,其理由和根
据又是什么?多谢答复! |
|
a******1 发帖数: 1031 | 6 线性模型可以看成是parameters向量和数据向量的linear combination,所以才叫“
linear” model. 如果模型的形式是b^(x),这个写不成b 和 x 的linear combination
,就成为non linear model |
|
t**********y 发帖数: 374 | 7 I did some test runs using built-in dataset and got different outputs:
1. ===
> test=lmer(Reaction~Days+(Days|Subject),sleepstudy)
> summary(test)
Linear mixed model fit by REML ['lmerMod']
Formula: Reaction ~ Days + (Days | Subject)
Data: sleepstudy
REML criterion at convergence: 1743.6
Scaled residuals:
Min 1Q Median 3Q Max
-3.9536 -0.4634 0.0231 0.4634 5.1793
Random effects:
Groups Name Variance Std.Dev. Corr
Subject (Intercept) 612.09 24.740
... 阅读全帖 |
|
x****h 发帖数: 78 | 8 【 以下文字转载自 Statistics 讨论区 】
发信人: xyzhhh (xiaoyuanzi), 信区: Statistics
标 题: 问一道multiple linear regression的题
发信站: BBS 未名空间站 (Wed Nov 30 22:11:20 2011, 美东)
Simple linear regression S1: Y~x(1) (i.e., Y=a1*x(1)+a0); R2 (coefficient of
determination) = 0.01 ;
Simple linear regression S2: Y~x(2) (i.e., Y=b1*x(2)+b0); R2 (coefficient of
determination) = 0.02 ;
Then multiple regression M3:Y~x(1) and x(2)(i.e., Y=c2*x(2)+c1*x(1)+c0); wha
t is the min and max of the R2 for this regression?
thanks! |
|
f*r 发帖数: 3968 | 9 LINEAR彗星(C2000 WM1)粉墨登场
译:Shea
在11月份,LINEAR彗星会穿越英仙座,用双筒镜也能轻松的找到它。在月底,说不定用肉
眼也能勉强看见它。南半球的观测者在12月份会有一个比较好的观测条件。
尽管这颗彗星在去年12月份被发现时只有18等,但它自那之后便开始慢慢的增亮。这儿的
星表是基于Syuichi Nakano提供的最新轨道根数计算的结果,其中给出了表中所给日期世
界时0时这颗彗星的赤经和赤纬,它和太阳的角距,目视亮度由John Bortle提供,以及它
所在的星座。
C/2000 WM1 (LINEAR) 轨道根数
过近日点时间 T:2002 Jan. 22.6747 TT
轨道偏心率 e:1.000281
近日点距 q: 0.555366 a.u.
近日点角距 ω:276.7703
轨道升交点黄经 Ω:237.8964
轨道面与黄道面的交角 i:72.5514
Epoch 2001 Oct. 18.0 TT = JDT 2452200.5
T 2002 Jan. 22.6747 TT Marsden
q 0.555366 (2000.0) P Q
z |
|
f*******0 发帖数: 14 | 10 有人说 linear 是 EE的基础,可是国内只有 搞 control 的才学 linear system,我
美国学校 EE各个方向第一门课都是
linear。要是没学好,是不是就废了。。。 |
|
E*****a 发帖数: 757 | 11 叫我说,linear system是个很重要的concept.必须吃透了。
这个是很多概念的来源,包括微积分吧。
很多东西想建立直觉,这个linear system是头一个需要搞得很清楚的东西
很多东西的初步分析,都是一个operation point+linear system analysis.
至于课程里边的什么方程之类,基本从来不用。 |
|
b****t 发帖数: 114 | 12 Hello all,
I am thinking about optimizing a piecewise linear function. Since the
explicit function form is too complicated, I only know the function form
piecewisely, i.e. a linear function for a small subset of the domain (e.g.
points in a unit d-dimensional simplex). Thus the overall function is
piecewise linear over the R^d space with integer breaking points. ( I hope I
state the problem clearly here).
The steepest descent method can be used to find the optimum of the function
here? Dose the |
|
l*****a 发帖数: 119 | 13 Linear Programming: Foundations and Extensions (很多人推荐,适合本科生,比较
通俗,例子还挺多)
Linear Programming with MATLAB (MPS-SIAM Series on Optimization) Michael C
. Ferris, Olvi L. MangasarianStephen J. Wright (很新的一本书,包括了很多LP
的新发展)
D. Bertsimas and J. Tsitsiklis, Introduction to Linear Optimization, Athena
Scientific, 1997
S. J. Wright, Primal-Dual Interior-Point Methods, SIAM, 1997
PS: 不是做理论的话,似乎没必要看这些书,毕竟现在LP的solver很成熟了,绝大部
分问题都能解 |
|
n*****1 发帖数: 172 | 14 请问linear manifold的具体定义是什么? 我在网上google了一下, 也看了一些书, 好
像有两个不同的定义.
第一个是, given a system of vectors x, y, z,..., a linear manifold spanned
by x, y, z... is the set of all finite linear
combinations ax+by+cz+....
另一个定义是, 假设V是一个vector space, L is a non-empty subset of V, if
there exists a v \in V such that L+v is a
subspace of V, then L is a manifold
这两个定义好像不equivalent吧
另外, 第一个定义说的情况, 和subspace有什么不同呢?
谢谢! |
|
h*****u 发帖数: 204 | 15 1 有3个变量y,x1 and x2,
2 假如correlation 系数(y,x1)=a1>0 and correlation 系数(y,x2)=a2>0,
3 Do the multi linear regression y as a funtion x1 and x2, suppose we get y~
b1*x1+b2*x2, 但是这里b1<0, b2>0.
从2中,我们得知y和x1的关系是同向变化的, 但是由于Do the multi linear
regression,我们得到 y~ b1*x1+b2*x2, 如果固定x2不变的话,我们就得到了y与x1
的关系是反向的。
我的问题是: 有没有一个办法让multi linear regression的系数的符号跟
correlation 系数的符号是一致的? 谢谢 |
|
a*****i 发帖数: 1045 | 16 Linear指的是参数的Linear,比如,a0, a1, a2是linear,没有^2 |
|
s********1 发帖数: 235 | 17 【 以下文字转载自 Statistics 讨论区 】
发信人: someone111 (some), 信区: Statistics
标 题: 有什么模型能把linear regression model 和 time series model 合起来做prediction 吗?
发信站: BBS 未名空间站 (Wed Dec 9 17:00:04 2015, 美东)
有什么模型能把linear regression model 和 time series model 合起来做
prediction 吗?现在有一些数据,是一堆产品,他们有一些自身的属性的数据,如网
上的ratings 等,他们还有time series 的销售数据,每一种产品有month by month
的销售值,现在要用这些数据做predictive modeling, 想到的方法有对自身属性的数
据可以做linear regression, svm 这样的模型预测,对time series 的销售数据,可
以用 time series 的模型预测,有没有什么模型能把这两种模型结合起来,用一个模
型考虑两方面,进... 阅读全帖 |
|
s********1 发帖数: 235 | 18 【 以下文字转载自 Statistics 讨论区 】
发信人: someone111 (some), 信区: Statistics
标 题: 有什么模型能把linear regression model 和 time series model 合起来做prediction 吗?
发信站: BBS 未名空间站 (Wed Dec 9 17:00:04 2015, 美东)
有什么模型能把linear regression model 和 time series model 合起来做
prediction 吗?现在有一些数据,是一堆产品,他们有一些自身的属性的数据,如网
上的ratings 等,他们还有time series 的销售数据,每一种产品有month by month
的销售值,现在要用这些数据做predictive modeling, 想到的方法有对自身属性的数
据可以做linear regression, svm 这样的模型预测,对time series 的销售数据,可
以用 time series 的模型预测,有没有什么模型能把这两种模型结合起来,用一个模
型考虑两方面,进... 阅读全帖 |
|
s*********8 发帖数: 334 | 19 Saab has been on top of my shopping list for a new car for a while. since my
budget
was around 20k, it was impossible to get a new saab 93/95. So I focused on
new saab 92x, used off lease 93/95.
92x linear auto was out of stock in my area and the shopping for saab moved
to used
93/95. Checked craigslist, found two saabs were of interest. One was 2004
saab 93 linear auto, 26k miles, asking price was 14.9k. The other one was
2004 95 linear (2.3T) auto, 30k miles, asking price was 16.9k.
CAlled a |
|
f*r 发帖数: 3968 | 20 LINEAR彗星(C2000 WM1)粉墨登场
译:Shea
在11月份,LINEAR彗星会穿越英仙座,用双筒镜也能轻松的找到它。在月底,说不定用肉
眼也能勉强看见它。南半球的观测者在12月份会有一个比较好的观测条件。
尽管这颗彗星在去年12月份被发现时只有18等,但它自那之后便开始慢慢的增亮。这儿的
星表是基于Syuichi Nakano提供的最新轨道根数计算的结果,其中给出了表中所给日期世
界时0时这颗彗星的赤经和赤纬,它和太阳的角距,目视亮度由John Bortle提供,以及它
所在的星座。
C/2000 WM1 (LINEAR) 轨道根数
过近日点时间 T:2002 Jan. 22.6747 TT
轨道偏心率 e:1.000281
近日点距 q: 0.555366 a.u.
近日点角距 ω:276.7703
轨道升交点黄经 Ω:237.8964
轨道面与黄道面的交角 i:72.5514
Epoch 2001 Oct. 18.0 TT = JDT 2452200.5
T 2002 Jan. 22.6747 TT Marsden
q 0.555366 (2000.0) P Q
z |
|
p*****2 发帖数: 21240 | 21 我做一道题,linear能pass 50%的test case, binary 只能过10%。虽然写的罗嗦,但
是还是很奇怪为什么这样子。
我后来从后往前linear search, 还是这个结果。 |
|
D******6 发帖数: 6211 | 22 【 以下文字转载自 Statistics 讨论区 】
发信人: Doha2006 (花猫), 信区: Statistics
标 题: 请教matlab non-linear optimization的问题
发信站: BBS 未名空间站 (Fri Jan 7 23:39:25 2011, 美东)
我需要用一个non-linear optimization function来estimate一个model,近似于logit
model,用maximum likelihood estimation。
如果用maxlik,其实这是一个min函数,就是名字叫max,从这个网站下载的,http://www.spatial-econometrics.com/
结果有点奇怪,不是指我的model参数奇怪,而是maxlik输出的结果如下:
CONVERGENCE CRITERIA MET: Change in Objective Function
iteration function value dfunc
0 2301.06004... 阅读全帖 |
|
D******6 发帖数: 6211 | 23 【 以下文字转载自 Statistics 讨论区 】
发信人: Doha2006 (花猫), 信区: Statistics
标 题: 请教matlab non-linear optimization的问题
发信站: BBS 未名空间站 (Fri Jan 7 23:39:25 2011, 美东)
我需要用一个non-linear optimization function来estimate一个model,近似于logit
model,用maximum likelihood estimation。
如果用maxlik,其实这是一个min函数,就是名字叫max,从这个网站下载的,http://www.spatial-econometrics.com/
结果有点奇怪,不是指我的model参数奇怪,而是maxlik输出的结果如下:
CONVERGENCE CRITERIA MET: Change in Objective Function
iteration function value dfunc
0 2301.06004... 阅读全帖 |
|
D******6 发帖数: 6211 | 24 【 以下文字转载自 Statistics 讨论区 】
发信人: Doha2006 (花猫), 信区: Statistics
标 题: 请教matlab non-linear optimization的问题
发信站: BBS 未名空间站 (Fri Jan 7 23:39:25 2011, 美东)
我需要用一个non-linear optimization function来estimate一个model,近似于logit
model,用maximum likelihood estimation。
如果用maxlik,其实这是一个min函数,就是名字叫max,从这个网站下载的,http://www.spatial-econometrics.com/
结果有点奇怪,不是指我的model参数奇怪,而是maxlik输出的结果如下:
CONVERGENCE CRITERIA MET: Change in Objective Function
iteration function value dfunc
0 2301.06004... 阅读全帖 |
|
y***u 发帖数: 101 | 25 you can convert |.| < c into linear constraints.
|.| > c is not linear. |
|
r********t 发帖数: 41 | 26 【 以下文字转载自 Programming 讨论区 】
发信人: rumorhasit (Love life), 信区: Programming
标 题: MINOS linear programming
发信站: BBS 未名空间站 (Thu Apr 24 20:31:57 2008)
Hello,
I need to write some fortran code to call minoss.f, minos solver, to solve
some linear programming.
This is my very first time to call minoss. I am wondering whether anybody
can give me a sample code how this works. Thanks a lot! |
|
p*******y 发帖数: 133 | 27 【 以下文字转载自 Software 讨论区 】
发信人: pennyjudy (penny), 信区: Software
标 题: how to do weighted linear regression in Excel?
发信站: BBS 未名空间站 (Sat May 22 13:07:59 2010, 美东)
Hi
I am trying to create a excel sheet to do weighted linear regression. This
sheet is intended to be shared with other ppl as a template.
I used user defined function downloaded from http://www.adamslim.com/ModellingGuides/ModellingGuidesWLR.htm. However, when I copy the excel sheet to other drive or other computer, it always report |
|
w**********y 发帖数: 1691 | 28 likelihood
A personal subjective suggestion: simply devide your log likelihood by the
number of your data, then you will have the sense of the goodness of fit.
Mean error
training error and true (predictive) error
-I didn't know how people utilized "cross validation" with 'holdout' data,
until I worked in an insurance company. Theoratically, what they did is not
that perfect.
AIC, BIC
There is no big difference for model evaluation in theory between linear and
non-linear regression. Just harder |
|
h********0 发帖数: 944 | 29 【 以下文字转载自 Statistics 讨论区 】
发信人: huiqiang80 (gene), 信区: Statistics
标 题: 包子求linear – quadratic model问题解决
发信站: BBS 未名空间站 (Sun Dec 12 16:20:58 2010, 美东)
最近做了Radiation的实验,现在手上有Data, 查了文献,要用linear – quadratic
model 来拟合一个曲线,版上有哪位用过这个模型的能教教我,小弟包子答谢!有意请
站内短信或留言详谈,多谢! |
|
w**d 发帖数: 2334 | 30 【 以下文字转载自 Mathematics 讨论区 】
发信人: wind (one), 信区: Mathematics
标 题: Textbook for linear algebra?
发信站: BBS 未名空间站 (Sat Sep 29 21:43:44 2007), 转信
Anyone knows some good linear algebra textbook? It is
for senior undergraduate or 1st year graduate. I just briefly
went through the books by G Strang and CW Curtis. Seems both are
not very complete. |
|
r********t 发帖数: 41 | 31 【 以下文字转载自 Programming 讨论区 】
发信人: rumorhasit (Love life), 信区: Programming
标 题: MINOS linear programming
发信站: BBS 未名空间站 (Thu Apr 24 20:31:57 2008)
Hello,
I need to write some fortran code to call minoss.f, minos solver, to solve
some linear programming.
This is my very first time to call minoss. I am wondering whether anybody
can give me a sample code how this works. Thanks a lot! |
|
f*****0 发帖数: 489 | 32
the answer depends.
most bipolars are used for linear applications. For obvious reasons, they
cannot work in the saturation region. However, there are switching bipolars,
and emitter-switched bipolars as well as igbt that do work in their
saturation region.
mosfets are primarily for switching applications so they work mostly in
their saturation region. However, most lateral mosfets (older mosfets or
audio mosfets) work in their linear regions.
In the end, most devices can work comfortably in ei |
|
H****E 发帖数: 444 | 33 请问non linear circuit simulator里是不是不存在纯粹从频域进行得模拟?比如说对
于一般电
路HSPICE里有small signal analysis,无须跑transient Simulation直接从频域解电
路,是不是
到了non linear circuit里,那些频域的模拟其实也是和时域的模拟相结合跑的(比如
从时域做傅
里叶分析-->频域)?
门外汉刚开始接触这方面的东西,看得云里雾里,求高手指点一二,也许会对继续的学
习起很大
帮助,谢谢! |
|
a******e 发帖数: 331 | 34 What kind of non-linear circuit? Normally the special simulation engine are
used for RF simulation such as Mixer, Oscillators etc.
从频域进行 Harmonic balance simulation, it is solving KCL/KVL in frequency
domain, in hspice, you can use .hb .hbac etc and ADS is strong in this.
Aother type analysis for non-linear RF circuit is shooting newton, in hspice
use .sn .snosc(for oscillators) .snnoise etc., in Cadence use PAC, PSS(
periodical steady state). Golden gate is good too. Popular for switch cap
circ |
|
c****n 发帖数: 21367 | 35 谢谢
1. For general linear programming, is interior point method still the
fastest in term of time complexity? I've heard that Karmarkar's algorithm
was bounded by O(n3.5 L), any improvement afterwards?
2. Is the strong-polynomial algorithm for linear programming still a hot
research direction? how does experts in this area, such as you, perceive the
difficulty and existence of strong-polynomial algorithm?
3. How about the average time complexity? We knew that about Simplex method.
Is there any re |
|
g****t 发帖数: 31659 | 36 理论我不懂.
如果是实用问题的话,我的经验,
小规模的问题用单纯型.大规模的就用牛顿法.
这样可以通过tuning确保,你的原问题A,
和A+e的结果差不多.(e是小扰动)
这点基本的鲁棒性挺重要的.因为你建模不可能绝对准确.
其实我更prefer全部都用牛顿法.Karmarkar方法和单纯形
概念都太复杂,不够直观.学过我就忘了.
牛顿法我个人比较熟悉.各种情况我自己将会弄好各种变形.
谢谢
1. For general linear programming, is interior point method still the
fastest in term of time complexity? I've heard that Karmarkar's algorithm
was bounded by O(n3.5 L), any improvement afterwards?
2. Is the strong-polynomial algorithm for linear programming still a hot
research direction? how does |
|
g****t 发帖数: 31659 | 37 linear system啥意思? 我看得有点晕.好像在国内确实没学过.
谁说国内只有搞control的才学linear system? |
|
d******e 发帖数: 7844 | 38 顺便问一下,我没学过Linear Programming。
有人能告诉我,Quadratic Semi-Definite Programming和这Linear Programming的关系
么? |
|
d******e 发帖数: 7844 | 39 LDA一般情况下就是指Fisher Linear Discriminant.
对于机器学习,一般来说,Dimensionality reduction只是一个提取特征/Feature
Extracion的过程,然后可以配合各种不同的分类器/Classifier。
当然,统计课本里教的Fisher Linear Discriminant是包括分类过程的,其实就是FLD&
Nearest Centroid + L2 distance |
|
i*****l 发帖数: 50 | 40 题目是
Consider the problem of minimizing a cost function of the form c'x+f(d'x),
subject to the linear constraints Ax>=b, Here, d is a given vector and the
function f: R->R is as specified in the following figure. Provide a linear
programming formulation of this problem.
\ /
\ /
\_________/ x-axis
1 2
左边斜率-1, 右边2
3x |
|
h********0 发帖数: 944 | 41 【 以下文字转载自 Statistics 讨论区 】
发信人: huiqiang80 (gene), 信区: Statistics
标 题: 包子求linear – quadratic model问题解决
发信站: BBS 未名空间站 (Sun Dec 12 16:20:58 2010, 美东)
最近做了Radiation的实验,现在手上有Data, 查了文献,要用linear – quadratic
model 来拟合一个曲线,版上有哪位用过这个模型的能教教我,小弟包子答谢!有意请
站内短信或留言详谈,多谢! |
|
F**S 发帖数: 13 | 42 I am embarrassed to ask such elementary questions but they have bothered me
like a bone in my throat for quite a while. Please be so kind to enlighten
me...
The questions are centered closely about infinite dimensionality:
How exactly to see that a mapping space is infinite dimensional? To be
specific, take the simplest example of the vector space, C(R, R), of
continuous functions from R to R, where R stands for the space of real
numbers. The statement must be obvious since books I have consulte... 阅读全帖 |
|
B****n 发帖数: 11290 | 43 C(R,R)的basis問題: 必然牽扯到一點分析 不可能純粹在linear algebra裡面討論 因
為牽扯到了極限(任何一連續函數 可被basis裡的函數的linear combination所逼近)
任何一本高微的課本應該都有 不需要用到什麼topology
C(R,R)是infinite dimensional的問題: 你只需要證明span{x, x^2,...,x^n}是n維的
for any n
me |
|
p***e 发帖数: 472 | 44 我在做数据分析,数据已经采好了,我的数据分析是两层的hierarchical linear
model。根据已有的数据,我的几个假设有的成立,有的没有被证明。我想看一下假设
成立和不成立的原因。
1. post hoc power analysis和confidence interval哪一个更有效?
2. 对于hierarchical linear model analysis,哪里可以找到计算的公式?有什
么软件可以做这类分析计算么?
万分感谢回复! |
|
I****k 发帖数: 35 | 45 最近要面一个位置,specific里说是个linear analytic group withtin rates
department. 有人知道什么叫linear analytic啊? |
|
p***e 发帖数: 472 | 46 我在做数据分析,数据已经采好了,我的数据分析是两层的hierarchical linear
model。根据已有的数据,我的几个假设有的成立,有的没有被证明。我想看一下假设
成立和不成立的原因。
1. post hoc power analysis和confidence interval哪一个更有效?
2. 对于hierarchical linear model analysis,哪里可以找到计算的公式?有什
么软件可以做这类分析计算么?
万分感谢回复! |
|
J*****n 发帖数: 4859 | 47 I am doing some linear prog optimization in excel through VBA.
In some case, parameters reached 1000, meanwhile, MS Excel's SOLVER's max
para allowrance is just 200.
I got some open source called opensolver. However, it doesn't suggest how to
run everything in VBA (Coz I need repeatly run linear prog).
Any suggestion?
Thank you. |
|
G******2 发帖数: 579 | 48 【 以下文字转载自 Statistics 讨论区 】
发信人: Gigi2012 (Gigi), 信区: Statistics
标 题: 请问multi variate linear regression 选择risk factor 问题
发信站: BBS 未名空间站 (Thu Apr 9 23:51:47 2015, 美东)
问大牛一个basic 的问题。
在evaluate 一个 multi variate linear regression.
几个factor 同时被研究,individual correlation coefficient 都不大,只有0.3左
右,不知道为什么被选进了model.
然后每个factor 的coefficient 的R square 都小于0.5。
这个 model 是个好model 吗?
还是我看错地方了?
我的问题1。 不是应该选correlation coefficient 大的factor to enter the model?
还是p value 更重要?
R square 低于0.5 应该留在model 里吗?
选择factor ... 阅读全帖 |
|
l*****g 发帖数: 8 | 49 I like Gilbert Strang's book too. However, I think it's more
directed
towards applied math, where things are mostly covered in
n-dimensional
Euclidean space. Therefore, some interesting things are left
out.
For example, the concept of group and field, abstract vector
space, linear functional etc.
If you want to know a little bit more than Euclidean space,
I would recommand
2 other books:
1) Paul Halmos: Finite Dimensional Vector Space
2) Morton Curtis: Abstract Linear Algebra
Another book coming |
|
d**********l 发帖数: 183 | 50 如果covariate里有time series 的regressor, response也是time series。这样的情
况下做linear regression是不是有些复杂?就我目前的理解是如果x和y都是
stationary的过程,或者如果x和y是cointegrated的,是可以直接做linear
regression只不过error的pattern 是stochastic的。
请问牛人们我的理解对吗?
如果y是stationary, x是i(1)那么是不是要把x difference 一次再对y 和diff(x)
做regression呢?
谢谢大虾的回答。 |
|