D******n
发帖数: 2836 | 1 it actually doesnt quite make sense to me to compare the coefficients of two
unrelated variables. It is more seen to compare the coefficents of the same
varialbe under different treatments.
y=b0+b1x1+b2x2+b3x1x2+e
x2 is class varialbe = 1 if in class 1, =0 in class 2
test if b3=0 |
|
S*******0
发帖数: 198 | 2 嗯，对的，varialbes in Java 会初始化default value。
谢谢。 |
|
w*****g
发帖数: 186 | 3 能多指导一下吗？ 我用==只是个例子，我想知道的是：
（1）方不方便得到dow的实时数据？
比如“dow>=71000",这样的条件在IB容易programing吗？
（2）在IB里面容易定义boolean varialbe吗？
比如：Boolean toBuy
if(toBuy)
buy GE 100
（3）容易得到time variable吗？
比如说我要在股市关门之前close my position
比如：if (currentTime >= 3:50pm)
sell my position |
|
n*****s
发帖数: 10232 | 4 【 以下文字转载自 Statistics 讨论区 】
发信人: waiting (让生命去等候), 信区: Statistics
标 题: 问个统计问题
发信站: BBS 未名空间站 (Wed Mar 28 12:34:26 2007), 站内
发信人: waiting (让生命去等候), 信区: Mathematics
标 题: 问个统计问题
发信站: BBS 未名空间站 (Wed Mar 28 12:32:41 2007), 转信
Two independently and uniformly distributed varialbes：
X1~U(a, a+1), X2~U(b,b+1), where a+1>b>=a>=0
Define X=max{X_1, X_2}.
What is the distribution of X?
What is the conditional distribution of X, on X
What is the mean of X, conditional on X
Thanks! |
|
l**********t
发帖数: 5754 | 5 In my belief, "random" doesn't exist. It is a belief so I can't prove it (or
can you prove it?). So I don't have a rigorous definition of "random"
aside from the textbook conceptual def. Even to your
virutal world exposured to the pseudo random numbers, I guess its
sample distribution can be inferred and converge to the true distribution as sample size increases (so there is a law associated with that pseduo "random" varialbe).
IMHO, for all practical purpose (as in Darwin's Evolution Theory),... 阅读全帖 |
|
T*******r
发帖数: 333 | 6 终于发现你的问题了。
第一点，变异是随机的，或者说intractable的，我可以同意。
第二点，自然选择为什么也是随机的？自然选择根本不是随机的。你不会把自然环境的
多样性和随机搞混了吧。
给定一个自然环境，如“沙漠”，“你们家后院的小池塘”或者“实验室的培养皿”，
把生物放进去之后的生存压力是确定的，而不是随机的。
举个最简单的例子，狮子对于羚羊来说就是自然选择的生存压力。这个压力是有固定方
向的：狮子“选择”跑的快的羚羊（让跑得快的生存下来）。狮子不会愚蠢到这一代的
时候专门吃跑的慢的羚羊，下一代的时候吃跑的快的羚羊，这一代吃母羊，下一代吃小
羊。这根本就不是一个随机过程。为什么自然选择在你看来会是随机的？就拿被狮子追
的羚羊来说，自然选择会让它“随机”的一代代的变得跑的快或者跑的慢吗？
or
as sample size increases (so there is a law associated with that pseduo "
random" varialbe). |
|
S**H
发帖数: 1256 | 7 Thanks!!!! It seems work now.
what I did
IN " Parameter Mapping"
new variable name: user::currentFile, ParameterName:0
type: varchar.
the sql code is like this now:
insert into dbo.logTable values
(?,getdate(),getdate())
*****************
Another Question:
If I have multiple varialbes needed to be passed.
How does the query know which ? is which ?
thanks !!!! |
|
y****9
发帖数: 144 | 8 这本质上是一个scalability的问题。所以首先要搞清楚到底是什么影响了scalablity.
没有具体的数据 (i.e metrics to point out which area is the bottleneck)。这
个问题只能对怎样提高数据库的scalability作些泛泛之谈。 for example, if it is
CPU-bound then it is relatively easy to scale by adding more CPUs ( in
Oracle RAC you can add another node), if it is I/O bound then it is more
difficult to scale usually.
Also we need to assume app design is good enough ( 90% time performance
problem is poor app design in my working env). If not, some areas such as
reduce un... 阅读全帖 |
|
m******t
发帖数: 2416 | 9
I am with you on this one - static varialbe is a bad idea.
But you haven't responded to my last post on why filter
doesn't help avoiding the parameter evaluation penalty. 8-) |
|
b******d
发帖数: 794 | 10 谢谢，我编程肯定不会搞什么public variable,但是我准备考scjp, 想搞明白点。
child不能access parent's private/default variable这个我知道，不过貌似系统开辟
child object内存的时候，给他所有parent的varialbe都开了空间，否则不能解释以上
现象: 比如改reference type就可以access parent's variable
另外一个类似的情况是child可以通过继承的public method access parent's private
variable也说明系统给parent's variable开辟了内存空间，并且初始化了 |
|
J*****n
发帖数: 4859 | 11 想实现boss worker model,但不想手动再写那些conditional varialbe之类的东西。想
找个现成的类。
谢谢。 |
|
g*****1
发帖数: 998 | 12 一个大概200M的csv file，几百万行，
每行格式一样，我用stringstream和getline读每一行，然后把这行信息(几个varialbe
)存入一个自己写的class object,然后把它push到stl的priority queue,然后做点小计
算。
现在的问题是太慢了，好几个小时没出结果。 请问有哪些地方我做的不对或者要改进
的，让这个过程快一些？ |
|
j*a
发帖数: 14423 | 13 priority queue这里慢了吧，200m个值啊
varialbe |
|
X****r
发帖数: 3557 | 14 The standard answer is to run a profiler on it (either
with a smaller data set or make it terminate earlier),
then you know what takes all these time.
varialbe |
|
S*********g
发帖数: 5298 | 15 你觉得很大的可能是，你push到priority queue里的时候花了大量的时间
push的complexity是 O(log n)
你花的总时间是 sum _{n=1}^N log(n) ~ N log N - N
如果每一步你花1ms,那么100万个记录对应的是 13815秒 = 3.8小时
varialbe |
|
g*****g
发帖数: 34805 | 16 对C++的IO类不熟，但直觉就是你的buffering太小。开个大缓冲区，
把东西读进来再处理，会好很多。如果读一行disk IO一次，就会很慢。
varialbe |
|
y**b
发帖数: 10166 | 17 使用流迭代器是不是会快一些？通常一个一个push进容器的效率不如给出一个迭代器范
围。
ifstream in("your_file");
istream_iterator in_iter(in);
istream_iterator eof;
priority_queue pq(in_iter, eof);
当然your_class需要重载输入操作符>>
varialbe |
|
r********3
发帖数: 2998 | 18 用buffered来读。 才200M的csv文件，以现在的硬盘速度，应该在10-20秒内就可以读
完并存入内存对象。
varialbe |
|
D*******a
发帖数: 3688 | 19 maybe faster to read everything and sort. 200m fits in memory easily.
varialbe |
|
n****n
发帖数: 568 | 20 我们写过类似的project,文件比这个还要大，全部读进来也就几分钟。你试着把文件先
全部读近。
varialbe
function,都是用来获取data member数值的。
oneOrder &b)，这个里面大量的用到了上面的2个member function |
|
a*********a
发帖数: 22 | 21 This question may be too specific, but --
What is @sys?
My path is:
>echo $PATH
/dept/union/csa/chinese/bin/@sys
(in the path of /dept/union/csa/chinese/bin,
there are subdirectories containing platform
dependent CXterm binaries, i.e, sun, sgi,
aix ans so on).
When I log on the system (maybe any one of
sun, sgi or aix workstation), the
/dept/union/csa/chinese/bin/@sys
will point to the corresponding CXterm
subdirectory (i.e., /dept/union/csa/chinese/bin/sun
if I log on a sun workstation).
T |
|
j***y
发帖数: 87 | 22 export PATH=/dept/..../`uname` |
|
z*****u
发帖数: 6 | 23 @sys automatically evaluates to the system id (e.g. sgi_65,
alpha_dux40..) under afs. it's file system specific and won't
work without afs. |
|
f***d
发帖数: 101 | 24 divided by zero, inexact, overflow, underflow, invalid
If you want your code to run continuously without exiting
for this error, your compiler should have one option for
that, for Sun f77, it is the default.
With deguggers like dbx, you will find the exact location
of such error in your code and you will be able to see values
of some varialbes in stack.
Good luck, |
|
f********8
发帖数: 5601 | 25 The difference is actually Censoring versus Truncation. Censoring means
corner solutions, the corresponding latent varialbes are negative.
Truncation simply means missing data, no observations are available, so
people put 0 when they collect data. Tobit is suitable for censoring, not
truncation. |
|
B****n
发帖数: 11290 | 26 嚴格來講是不對的 但一般情況是對的
因為按照定義 continuous random varialbe只是指distribution滿足一些性質 (比如
說CDF is absolute continuous) 但並沒有說對每一個w (point in sample space) 是
如何
point |
|
|
w*****g
发帖数: 47 | 28 【 以下文字转载自 Statistics 讨论区 】
发信人: waiting (让生命去等候), 信区: Statistics
标 题: 问个统计问题
发信站: BBS 未名空间站 (Wed Mar 28 12:34:26 2007), 站内
发信人: waiting (让生命去等候), 信区: Mathematics
标 题: 问个统计问题
发信站: BBS 未名空间站 (Wed Mar 28 12:32:41 2007), 转信
Two independently and uniformly distributed varialbes：
X1~U(a, a+1), X2~U(b,b+1), where a+1>b>=a>=0
Define X=max{X_1, X_2}.
What is the distribution of X?
What is the conditional distribution of X, on X
What is the mean of X, conditional on X
Thanks! |
|
p********a
发帖数: 5352 | 29 ☆─────────────────────────────────────☆
pome (穿着一件花衣裳象一朵秋海棠) 于 (Fri Oct 19 13:33:13 2007) 提到:
比较艰难的过了。
1，有一个问为什么要hash object？ 以前一直没听说过。。
2，infile时， filevar= ka
3, infile
tnnd, www下发文不让修改？
Advanced techniques
☆─────────────────────────────────────☆
pome (穿着一件花衣裳象一朵秋海棠) 于 (Fri Oct 19 13:34:33 2007) 提到:
以前置顶的130题， 因为重复太多， 大概考了10道。。
☆─────────────────────────────────────☆
pome (穿着一件花衣裳象一朵秋海棠) 于 (Fri Oct 19 13:35:42 2007) 提到:
5个& resolve一个macro varialbe。
☆──────────────────────────── |
|
a********a
发帖数: 346 | 30 Is there anyway if I only treat x and y as one varialbe to get overall mean
of x and y, but treat z seperately to get mean of z only?
Thanks |
|
A*********u
发帖数: 8976 | 31 options mprint mlogic symbolgen;
you can also use
%put phat=&phat;
to check the value of macro varialbe at any place
编了一段很长的macro来实现某个复杂模型下的EM algorithm,里面有3个macro
variable分别是&phat, &muhat, &sigma
我设定程序循环10次，所以应该产生10组phat, muhat,sigma
可是我不知道怎么把这10组phat, muhat,sigma在每一次循环的结尾打印出来，或者最
后产生一张表打印出所有10组phat, muhat,sigma。有高手知道怎么写SAS code么？不
胜感激。。 |
|
z*****e
发帖数: 5 | 32 例如：有一个dataset with three observations as:
var1
a
b
c
怎样多加一个变量使这个dataset变成
var1 var2
a a
b ab
c abc
谢谢！ |
|
p********a
发帖数: 5352 | 33 data b;
set a;
if _N_=1 then var2=Var1;
else Var2=cats(lag(var2),var1);
run; |
|
|
|
p********a
发帖数: 5352 | 36 这句哥哥把俺叫得心惊肉跳的，俺又不是那个同同，哥哥不是能随便叫的。和俺学不如
找书学。 |
|
l*****i
发帖数: 43 | 37 data a;
input var1 $;
cards;
a
b
c
;run;
data b;
set a;
retain y;
length y $ 8;
y=trim(left(y))||' '||var1;
run; |
|
|
s*********o
发帖数: 42 | 39 如果要比较varialbe prog 的level 2 vs. level 3, 以及group 1 和 group 2 之间的
区别。prog 一共有3个level.
sas code如下：
proc logistic data = hsb2;
class prog/param=glm;
model hiwrite(event="1")=female read math prog;
contrast '1 vs 2 of prog' prog 1 -1 0/estimate;
run;
请问这里是怎样给prog做coding 的，为什么prog是1 -1 0, 为什么 contrast '1 vs 2
of prog' ,给出来的是1 vs 3 and 2 vs 3的odds ratio estimate？ 谢谢。 |
|
l*****k
发帖数: 587 | 40 I guess you can write an easy function that takes arguement
like CA, AC or FL, subset it in function then do the plot
you can make your funciton accepts multiple arguments so you can
do more complex subset and plotting.
you can also let your program iterate through all the varialbes,
plot to pdf or jpg file, so it runs automatically. |
|
d******r
发帖数: 1389 | 41 descending只对紧跟着的variable有用
这里的month是character varialbe，所以按字母顺序 |
|
s******o
发帖数: 283 | 42 Thanks so much for your help again, dido2009.
I know you did not sort the data set A before you use it for data B.But
does
data set A itself look like it is sorted already? because A variable in
this dataset is arranged in ascending order(1,5,7,9) already? and when
you
use it for data B, there is
set A;
By A;
with the combination of "set" and "by" statement, it require that data A
is
sorted or indexed as I referred above in SAS base"By group processing"
??
Do you know the result if the data set ... 阅读全帖 |
|
i*********e
发帖数: 783 | 43 If I have a data: 1500 records: 50 records with the depenedent varialbe's
vlaue of 1.
In this condition, if I use spss, should I just select only 50 records with
the dependent varaible's value of 0 and put the other 50 records with the
dependent variable's value of 1 together.
SPSS has no function of K fold cross validation, as far as I know.
Thanks! |
|
i**s
发帖数: 332 | 44 计算weighted average X weighted by Y = sum(Xi*Yi)/sum(Yi)
那么当x weighted by two varialbes Y and Z，如何计算？ 比如一门课的平均分X，
有几个不同阶段的小考分数Xi，每阶段的小考有自己的权重Y，同时又需要考虑出勤课Z
时的权重，怎么算 - 是把Y*Z当成权重吗？ 多谢！ |
|
E**********e
发帖数: 1736 | 45 multinomial logistic regression. each question is a dummy varialbes.
response variable is > 3 levels. if there is no response variable, how can
you do regression? unsupervised clustering? |
|
E**********e
发帖数: 1736 | 46 我的问题是用pca 来选signifiant variabales。 好像还没看到pca可以适合
categorial variables的。 昨天问了老师，答案也是negative， 倒是提了
correspondse analysis。 不过不是用来variables selections的。
比如一个binary variable（0,1）， 难道可以用PCA来算它的variance的大小， 从而
确定看在total variance占的比例大小决定保留还是剔除。
我的一个想法是把categorial 和continouous 分开， 分开分析从而确定significant
varialbes
interpret |
|
r**********9
发帖数: 19633 | 47
哇。。这种社会统计学 MODEL是极其复杂的。。。很难用矩阵运算
你的意思是说。。通过 ITERATIVE 得的方法，算 MODEL COVARIANCE 矩阵？
可是好几个 VARIALBE是 LATENT的。。。呜呜呜。。。这个可行吗 ？ |
|
y*******n
发帖数: 10103 | 48 不知道有没有varialbe valve timing。 |
|
l*********y
发帖数: 3447 | 49 what do you mean by two events correlated? if you mean two random variable,
then what you mean by two random varialbe intersect?
do not understand the question... |
|
