|
|
v*****u
发帖数: 1796 | 3 you're right. the algorith seems to have such limitation. The reason behind
is that they're not reducing for the same node. |
|
|
|
e*******n
发帖数: 872 | 6 请有兴趣的同学吧你的名字，公司，职位，邮箱通过站内邮箱发给我
谢谢！ |
|
s******y
发帖数: 163 | 7 课本是被人当成垃圾遗弃的，
Amazon链接
http://www.amazon.com/A-Friendly-Introduction-Numerical-Analysi
http://www.amazon.com/Operations-Research-Applications-Algorith
通俗小说是中文的，质量还不错，谁要消遣谁拿走
联系邮箱[email protected]
(function(){try{var s,a,i,j,r,c,l,b=document.getElementsByTagName("script");l=b[b.length-1].previousSibling;a=l.getAttribute('data-cfemail');if(a){s='';r=parseInt(a.substr(0,2),16);for(j=2;a.length-j;j+=2){c=parseInt(a.substr(j,2),16)^r;s+=String.fromCharCode(c);}s=document.createTextNode(s);l.parentNod... 阅读全帖 |
|
|
M*********e
发帖数: 1988 | 9 if you are asking a research question, then check papers on streaming algorith
ms that uses a log space to do approximate counting. |
|
b***y
发帖数: 2799 | 10 ☆─────────────────────────────────────☆
kevinyue (飓风) 于 (Tue Oct 11 16:15:03 2005) 提到:
正在把自己从前的一个matlab 程序改写为C code,因为多处用到hilbert(), 而C又没有这
个自带函数, 敢问哪位能提供一个C code for Hilbert Transform. 多谢多谢。
☆─────────────────────────────────────☆
alexx (panda in love~八胖~饲羊员~水木十年) 于 (Tue Oct 11 16:19:06 2005) 提到:
google ya.
http://www.cs.sunysb.edu/~algorith/implement/FFT/distrib/
这
☆─────────────────────────────────────☆
thrust (Thrust Jaeina) 于 (Tue Oct 11 16:34:00 2005) 提到:
hehe, matlab |
|
k*******d
发帖数: 1340 | 11 I understand, because you were in math. The proofs are in some algorithm
books like the mit one, or algorith design by ppl from Cornell. |
|
B***i
发帖数: 724 | 12 建议楼主看一下local alignment ( smith-waterman algorith) 吧, |
|
T*******n
发帖数: 493 | 13 What do you not like about \hat? The alignment or the
shape of the printed hat?
Assuming you are using LaTeX, try \usepackage{amsmath}
which uses a different algorith for aligning the hat.
Or try the accents package, found at
http://www.ctan.org/tex-archive/macros/latex/contrib/bezos/ |
|
s*y
发帖数: 472 | 14 自从两年前作出转行的决定以后，就没怎么上来灌水了。下午拿到Bloomberg的Financi
al Application Engineer offer，虽然意料之中，但是回想这两年还是有些感慨，上来
唠叨一下。
6年前，怀着一腔对科学的热情，据掉USC的Molecular and Computation Biology的$28
,000一年的offer（总觉得计算生物学不太像是严谨的科学），来到PSU搞生物。但越是
读下去，越是发现自己不喜欢也不适合做生物。对事业和生活的热情和憧憬不断被消磨
，整个人的状态越来越差。当年一心认定做科研的我迷茫，无助，不知该往哪里走。曾
经想过转生物统计PhD，无奈PSU没有这个专业，那时候特热门的统计又不太喜欢。于是
在彷徨和犹豫中，走过了三年半。
一个偶然的机会，让我选择了Engineering。做了一些research，问了一些人以后，选择
了Computer Science。于是新学期开始后，跟老板坦白，然后边做TA，边做实验，边去
计算机系旁听课程。记得那时候最后坚持下来的就是本科四年级的Data Structure and
Algorith |
|
h*******u
发帖数: 142 | 15 try google National Weather Srevice's NEXRAD weather radar, both 5min and 15
min are available.
However, radar rainfall data is not reliable. You need extra rain gauge plus
algorith to get a sound result.
good luck |
|
f***a
发帖数: 329 | 16 你是要求积分还是要generate samples from distribution？
generate samples from distribution的话：有buit-in funtion的就用，不然有很多别
的办法，譬如inverse CDF, transformation, accept-reject algorith, important
sampling, ..., etc. MCMC是其中一种，一般来说MCMC运算量大，不是特别麻烦的
distribution,能用别的algorithm就用别的吧。
你是要求f(x)积分的话：常见的当然直接用CDF就完了；f(x)是很复杂的pdf，或者是很
复杂的
f(x)的函数的话有很多numerical methods可以用（具体看相关的书），其中一种方法
就是用
distribution sampling求，譬如
\integral f(x) dx= \integral f(x)/g(x)*g(x ) dx = E[f(x)/g(x)];
generate x_i, i=1,...,n from g(x);
then the e |
|
A********r
发帖数: 43 | 17 For Stat people, is class like data structure and algorithm necessary?
seems data structure and algorithm are required for taking machine learning
also I'm not sure if it could be handled. |
|
m*********n
发帖数: 413 | 18 我觉得有帮助，虽然很具体的数据结构或者算法在统计中（尤其工作中）往往用不上，
但这方面的知识对统计中的coding efficiency很有潜在的帮助。
learning |
|
w**********y
发帖数: 1691 | 19 seems data structure and algorithm are required for taking machine learning
-- I don't think that is necessary, but totally depends on the course plan.
Take a look at the material here:
http://ocw.mit.edu/courses/electrical-engineering-and-computer-
You will then get a sense whether you can handle it.
learning |
|
m*********n
发帖数: 413 | 20 agree。statistical machine learning 不太需要算法和数据结构。
有些可能要，比如各种network models，但也是简单需要知道怎么represent a graph
structure, 和一些简单的algorithm，比如广度优先和深度优先。这些不上课，看看书
就能懂（我假定广度优先和深度优先这种能自己看懂编出来）。
learning |
|
A********r
发帖数: 43 | 21 thanks a lot for all your comments and suggestions.
statistical machine learning 不太需要算法和数据结构。---> the machine
learning class I'm planning to take is offered by computer science
department. that's why I'm sort of worried.
I'll take a look at that online course instruction and make a decision |
|
w**********y
发帖数: 1691 | 22 ask for a syllabus for that machine learning course too...and then maybe
paste it here..lol
I took that course from both CS and Stat. The only difference i can feel is,
statistical learning course is more like math flavor with lots of theories
and formulas; ML course is closer to engineering flavor with more pictures
and intuitive explanations. |
|
c*****a
发帖数: 808 | 23 data structure and algorithm 没什么用吧。就是教你cs里面的常用架构coding多。
里面的算法应该是graph theory, tree,NP问题多吧
我学校的ML require是AI |
|
B******5
发帖数: 4676 | 24 呃，我们也是require AI，不过似乎没啥需要的地方。。。 |
|
d********h
发帖数: 2048 | 25 AI不需要太多算法，到是data struc需要很强的编程能力，是我迄今为止最为痛苦的一
门课，有是熬夜作业还没写完 |
|
t******q
发帖数: 117 | 26 【 以下文字转载自 EE 讨论区,原文如下 】
发信人: TigerJade (小人莫不有才), 信区: EE
标 题: Re: 请问DCT可用于IMAGE COMPRESSION的关键原因是什么？
发信站: Unknown Space (Thu Aug 21 14:27:08 2003) WWW-POST
I agree with your first 2 points. But research work cannot be limited by the
standard. Who can know the hardware complexity even before the algorithm has
ever been proposed? At the same time, the complexity can be supported by the
hardware is increasing. In the research, we are focusing on the futures
application. No one can say is a algorith |
|
