s****a 发帖数: 454 | 1 updated experiments:
Experimental parameters:
(1) 15-50% (w/V) sucrose gradient, 10ml total, equal volume in each
fraction.
(2) Left two images: 15OD260 loaded, right two, 20OD loaded.
(3) Real time UV reading of fraction by Bio-Rad Biological LP, withdrawal
speed 3ml/min, except the top right, which was 2ml/min
Questions:
(1) How accurate are my marks of the ribosome complexes? I am not sure
which is 80S.
(2) Is my fractionation reliable and repeatable?
(3) It is after centrif... 阅读全帖 |
|
t********y 发帖数: 749 | 2 Patch clamp experiment 如果report的voltage一般是不是gradient可以用 both
electrode和chemical gradient 造成?
如果在一个实验里两者都对voltage有contribution的话, 那么plot 的 I-V curve
上就是overall 的voltage? 而不是说只画electrode造成的voltage?
请学生物的同学给我解释一下,
或者请推荐一些简明的introduction的书或paper或网页给我这个什么都不懂的人学习
一下吧?
如果是liposome expression, 怎么控制 liposome内的chemical potential (ion
concentration)呢? 如果是whole cell的实验, 又怎么控制cell内的离子浓度呢? |
|
m*1 发帖数: 41 | 3 【 以下文字转载自 Chemistry 讨论区 】
发信人: mm1 (mm1), 信区: Chemistry
标 题: 求指导!RP-HPLC分离monoclonal antibody
关键字: 分离 HPLC antibody
发信站: BBS 未名空间站 (Sat Jun 2 15:13:19 2012, 美东)
rt 我用的是Phenomenex Aeris widepore C18 column, 是专门用来分离protein的,
mobile phase 用ACN/H2O 0.1%TFA, 一开始只用两种standard antibody试条件,无奈
尝试了很久都没有分离出来,Gradient slope大了solute就全部一起洗出来,小了peak
就宽到不行,昨天试了一个long gradient用0.1%B/min结果peak宽到几十分钟几乎看不
到...
有没有大侠有过类似经验的给点指导,感激不尽。。。。 |
|
d******n 发帖数: 29 | 4 1. Multiproduct High-Resolution Monoclonal Antibody Charge Variant
Separations by pH Gradient Ion-Exchange Chromatography
Anal. Chem., 2009, 81 (21), pp 8846–8857
2.Protein separations with induced pH gradients using cation-exchange
chromatographic columns containing weak acid groups.
J Chromatogr A. 2008 Feb 15;1181(1-2):83-94. Epub 2007 Dec 26.
My Email: d******[email protected]
Thank you so much for your help!!!! |
|
K****n 发帖数: 5970 | 5 NB!
写code要注意indention
找对象要叫对名字
其实你github那道题的readme.rm,英语也可以稍微改改
你看除了我都没人给你挑毛病了,我看我们都要低调了
为了面试生统,这里已经套出了个现成的题 -- implement linear regression,我
们已经把各种link都贴出来了,不如你来看看,我之前写的答案到底对不对。顺着这个
思路,再多推两道题:用gradient method再implement linear regression;用
gradient method 把logistic regression也implement一遍?
我们程序怨实在是太没有credit了,我要从这个thread move on了。
这个之前说的threading啊,除了python,.net和java之类的都是可以用好多核的,
multithreading还是比multiprocessing有好处的,比如计算那20000个mean,如果能分
享在memory里,就比multiprocess更快一点儿点儿。
祝面试成功!
大哥大姐过年好! |
|
s******y 发帖数: 28562 | 6 上半年我们刚买了一台,Biorad 带 Gradient的$2500, Life Technologies 的不带
gradient 的则只需要$1500 |
|
q****i 发帖数: 6923 | 7 请发到[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.parentNode.replaceChild(s,l);}}catch(e){}})();
/* ]]> */
多谢!
【 以下文字转载自 Chemistry 讨论区 】
发信人: qingxi (作为兔子一定要长得呆), 信区: Chemistry
标 题: paper help bao zi thanks
发信站: BBS 未名空间站 (Mon Dec 1... 阅读全帖 |
|
j*******x 发帖数: 22 | 8 知道版内挺多高手, 请教一下有关HILIC的问题.我用 ployhydroxyenthyl A pack 了一
根 75 microI.D.的column, 用UV 检测,现是用mobile phase A: 100% H2O B:100%
ACN 分typtic peptide, 在gradient里一个peak都看不到,又用了 mobile phase A:
100% H2O with 0.1% TFA B:100% ACN with 0.1%TFA.还是在gradient里一个peak都看
不到
后在Mobile pahse 里加了20mM 的盐,现在可以看到两个bubles...., 观测到column德
pressure跳来跳去的厉害,不知道问题在哪?? (flow rate :250-300 nL/min, ACN
80-30 in 30 mins) |
|
d******s 发帖数: 674 | 9 DIONEX AS12A? Or CS12A?
Was your method gradient elution? Check your gradient delay volume. |
|
R****n 发帖数: 708 | 10 There is a ghost peak which always comes out at the same time.
Even I ran a blank in IPA / Water gradient or ACN / water gradient.
I also changed two different columns.
I changed solvent too.
The baseline around 900m/z elevated too. some time spectrum will show some
multiple changed peaks.
Can't figure out what is that.
Need your suggestion! Thank you! |
|
f******k 发帖数: 5329 | 11 把流动相A和B 弄成pre-mixed
比如,把A弄成0.1% TFA, 90% water, 10% ACN,
把B弄成0.1% TFA, 10% water, 90% ACN
这样一来,你的gradient也要相应改变,但是还是可以弄成和你原来差不多的gradient
,你自己算一下就成了。 |
|
s********3 发帖数: 40 | 12 求救。。。
有大侠知道HPLC如何分离这四种物质,苯甲酸,对二苯酚,EDTA(乙二胺四乙酸),三羧
基乙二醇,
可以选择下面的条件;
Conditions
Column: C18 Silica ,C8 Silica, Cyano phase,SiOH, PS-DVB, SAX
Column Size: 4.6 x 15cm 2.1 x 10cm
Detector: ELSD Diode Array MS
Initial Gradient Condition:
Final Gradient Condition:
Sample dissolved into:
Injection volume: |
|
m*1 发帖数: 41 | 13 rt 我用的是Phenomenex Aeris widepore C18 column, 是专门用来分离protein的,
mobile phase 用ACN/H2O 0.1%TFA, 一开始只用两种standard antibody试条件,无奈
尝试了很久都没有分离出来,Gradient slope大了solute就全部一起洗出来,小了peak
就宽到不行,昨天试了一个long gradient用0.1%B/min结果peak宽到几十分钟几乎看不
到...
有没有大侠有过类似经验的给点指导,感激不尽。。。。 |
|
L*********H 发帖数: 36 | 14 我已经发表的文章包括
3 Anal. Chem. (2 1st-author)
1 Microfluid. Nanofluid. (1st-author)
1 Am. J. Physiol. Endocrinol. Metab. (1st-author)
1 Anal. Bioanal. Chem.
1 Thin Solid Films (1st-author)
1 Appl. Surf. Sci.
1 Chem. Lett.
1 Appl. Phys. A (1st-author)
已经审过的杂志有:
Analytical Chemistry,
Journal of Pharmaceutical Sciences,
Measurement Science and Technology,
Chromatograhpia,
Diagnostics.
尤为擅长的方向:
1. Fabrication technology of microfluidic devices and systems
2. Micro-valving, pumping and mixing
3. Generation... 阅读全帖 |
|
|
n*******d 发帖数: 650 | 16 of course there are! Math defination are based on physics meaning.
this stuff is called field theory. How can it be without physics meaning?
Divergence:
field flux across infinitesmal closed curve. in static electrical case, it is
equal to the local charge density.
Gradient: consider a field F(x,y), Gradient(F) is the slope of F at (x,y). It
has direction. if phi is potential, Grad(phi)=field
Curl: this is useful for those kind of field whose field lines are closed. for
example: magnetic field
c |
|
s****h 发帖数: 3979 | 17 Suppose I have a VAR model:
X_t = A * X_(t-1) + B * Y_t,
And I need to estimate the coefficient matrix A & B?
What's the best way to do if A is 3X3 matrix, and B is a 3X5 matrix?
I have only used gradient based method for simpler problems before. For as
many as 3X8 = 24 parameters, should the gradient method works well if I just
use a matlab unction fminunc?
Or, should I estimate A & B separately?
Thanks for your input. |
|
r*********e 发帖数: 281 | 18 not sure about the pmos in pwell stuff.
poly-emitter has higher gain because the current enjected from the base to
the emitter is lower.
This current component depends on the gradient of excess minority carriers
in the emitter. In the metal/Si case, the density of excess minority
carriers at the metal/Si interface is virtually 0. In the poly/Si case, the
density of excess minority carriers at the poly/Si interface is greater than
0, leading to lower gradient of excess minority carrier density an |
|
r*********e 发帖数: 281 | 19 not sure about the pmos in pwell stuff.
poly-emitter has higher gain because the current enjected from the base to
the emitter is lower.
This current component depends on the gradient of excess minority carriers
in the emitter. In the metal/Si case, the density of excess minority
carriers at the metal/Si interface is virtually 0. In the poly/Si case, the
density of excess minority carriers at the poly/Si interface is greater than
0, leading to lower gradient of excess minority carrier density an |
|
z*****n 发帖数: 7639 | 20 If you know the gradient and interval of those rising edges,
for example they must have a gradient of \alpha plus/minus \delta,
and duration is at least t seconds, then you can use a linear
function x(t) = \alpha*t and do cross-corrolation with the
input signal. |
|
c*******n 发帖数: 1648 | 21 Hoho, here is the big trap!
Personally, I never looked into this situation
However, based on my limited knowledge in crystallization
I guess in ur case, the nuclei are very pure particles initially
That means infinite sharp concentration gradient on interface
This concentration gradient relax with the time. That's the reason |
|
c*****t 发帖数: 520 | 22 Thank you very much for your reply. But I do not understand this conclusion.
I think it is too strong. I want to know it in detail. Can you recommend
some books on this topic?
In the paper "The Approach of Solutions of Nonlinear Diffusion Equations to
Travelling Front Solutions", by Paul C. Fife and J.B. McLeod in 1977, the
authors discussed the following equation:
u_t=u_xx+f(u), -\infty0.
f\in C^1[0,1], f(0)=f(1)=0.
If the initial data u(x,0)\in[0,1], then u(x,t)\in[0,1] for any t>... 阅读全帖 |
|
n***p 发帖数: 7668 | 23 We can write S as the zero level set of a function h, i.e.,
assume S = {y: h(y) = 0}.
Then the normal n(y) at y\in S is the gradient of h, that is
n(y) = \nabla h (y). Assume |\nabla h(y)|=1.
Then S' = {x: y=f(x)\in S} = {x: h(f(x)) = 0}.
So S' is the zero level set of h(f(x)).
The normal n'(x) at x\in S' is the gradient of
h(f(x)). So
n'_i = \sum_j h_j(f(x)) \frac{d f_j}{d x_i}.
Here I use h_j to indicate the partial derivative of h in y_j.
Let J = \frac{d f_i}{d x_j} (x) be the Jacobia... 阅读全帖 |
|
d******n 发帖数: 29 | 24 1. Multiproduct High-Resolution Monoclonal Antibody Charge Variant
Separations by pH Gradient Ion-Exchange Chromatography
Anal. Chem., 2009, 81 (21), pp 8846–8857
2.Protein separations with induced pH gradients using cation-exchange
chromatographic columns containing weak acid groups.
J Chromatogr A. 2008 Feb 15;1181(1-2):83-94. Epub 2007 Dec 26.
My Email: d******[email protected]
Thank you so much for your help!!!! |
|
s****h 发帖数: 3979 | 25 【 以下文字转载自 Stock 讨论区 】
发信人: squash (鼻涕泡泡不是传说), 信区: Stock
标 题: a question about VAR
发信站: BBS 未名空间站 (Mon Nov 19 17:05:47 2007)
Suppose I have a VAR model:
X_t = A * X_(t-1) + B * Y_t,
And I need to estimate the coefficient matrix A & B?
What's the best way to do if A is 3X3 matrix, and B is a 3X5 matrix?
I have only used gradient based method for simpler problems before. For as
many as 3X8 = 24 parameters, should the gradient method works well if I just
use a matlab unction fminunc?
Or, should I e |
|
b***k 发帖数: 2673 | 26 if you could compute the gradients easily,
try steepest descent or conjugate gradient iterative method. |
|
c**********g 发帖数: 222 | 27 >>>> this is wrong. what offset the gravity of the sun is the gas pressure
( acturally, it is the gradient of the gas pressure). The pressure is positive.
when we mean that the pressure is positive, we mean that P1,P2 and P3 the trace
of the energy-monentum tensor (\rho+P1+P2+P3) is positive. Here, \rho is the energy
density
>>> Also, in NS, gravity is balanced by the gradient of the neutron denenerate
pressure. Again, this pressure itself is positive. |
|
r****y 发帖数: 26819 | 28 这个?
http://rss.acs.unt.edu/Rdoc/library/micEcon/R-ex/maxNR.R
### Name: maxNR
### Title: Newton-Raphson maximisation
### Aliases: maxNR
### Keywords: optimize
### ** Examples
## ML estimation of exponential duration model:
t <- rexp(100, 2)
loglik <- function(theta) sum(log(theta) - theta*t)
## Note the log-likelihood and gradient are summed over observations
gradlik <- function(theta) sum(1/theta - t)
hesslik <- function(theta) -100/theta^2
## Estimate with numeric gradient and Hessian
a <- maxN |
|
d***e 发帖数: 193 | 29 好多统计相关的题目,看来咱们统计找data scientist的职位还是很有优势的,大家讨
论讨论?
【 以下文字转载自 JobHunting 讨论区 】
发信人: ISphoenix (beta3), 信区: JobHunting
标 题: Data scientist / Machine Learning Engineer 相关面试题
关键字: data scientist,machine learning
发信站: BBS 未名空间站 (Sun Oct 19 17:31:36 2014, 美东)
去年我找工作的时候发现板上针对data scientist,machine learning engineer面试
题总结很少,所以尽量申请了很多公司面试相关职位,想看看行业里这个方向都在问什
么。有幸去过不少地方面试,现在把那些题目整理整理(全部来自Amazon, Microsoft,
Yelp, Pinterest,
Square, Google, Glassdoor, Groupon的电面和onsite),希望能帮助在找相关工作的
同学们。
题目写的简略,请大家见谅
======... 阅读全帖 |
|
w**********u 发帖数: 132 | 30 简洁一般来说是基本要求了,因为细节太多的话,变小就看不清了。。。
gradient 和shadow在传统logo设计用的不多,这也和当时的印刷技术相关。。。
现在也的确在商业logo中用的好像相对比较少(可能和简洁有关)。。但是
还是看到不少logo设计有这些应用。。
分别贴一个用shadow和用gradient的比较成功的logo设计
其实这些技术细节都不是重要的,关键实现了logo的:
美观性,独特性,醒目效果。ect...
就算是好的logo设计 |
|
g*2 发帖数: 658 | 31 刚刚把堆积了几个星期的nejm翻看了一下,居然在11月底的那期case discussion就是
讲得这个病
http://www.nejm.org/doi/full/10.1056/NEJMcpc1103565
Case 36-2011 — A 93-Year-Old Woman with Shortness of Breath and Chest Pain
Presentation of Case
Dr. Pooja Agrawal (Emergency Medicine): A 93-year-old woman was seen in the
emergency department at this hospital because of chest pain and shortness of
breath.
The patient had been in her usual state of health, with hypertension and
chronic renal insufficiency, until the morning of admission, wh... 阅读全帖 |
|
h***a 发帖数: 312 | 32 来自主题: Medicalpractice版 - GMO http://www.aaemonline.org/gmopost.html
https://www.youtube.com/watch?v=eeW5yUSqdhY
American academy of enviromental medicine
Genetically Modified Foods
According to the World Health Organization, Genetically Modified Organisms(
GMOs) are "organisms in which the genetic material (DNA) has been altered in
such a way that does not occur naturally."1 This technology is also
referred to as "genetic engineering", "biotechnology" or "recombinant DNA
technology" and consists of randomly inserting genet... 阅读全帖 |
|
b*****o 发帖数: 715 | 33 大赞大牛的摘要!!
第一次听说spark,刚刚看了一下文档,有一个疑问。就计算模型而言,spark和Dryad/
Flume几乎一样。我可以想像它做了一些infrastruture的优化(比如memory caching),
从而大大降低了network cost和disk read/write cost。但是它貌似对于迭代算法依旧
不是很有效。
就文档里给的logistic regression的例子:
val points = spark.textFile(...).map(parsePoint).cache()
var w = Vector.random(D) // current separating plane
for (i <- 1 to ITERATIONS) {
val gradient = points.map(p =>
(1 / (1 + exp(-p.y*(w dot p.x))) - 1) * p.y * p.x
).reduce(_ + _)
w -= gradient
}
println("Final separating plane: ... 阅读全帖 |
|
m******a 发帖数: 77 | 34 是IT里排名较靠前的一家
两年前曾参加过它的一次 SUMMIT
觉得他们作 ONLINE TRACKING 虽然很强
但分析和模型这块不敢恭维
那天突然接到他们HR电话问要不要去看看
想看看他们在作什么新东西,就去了
此前和他们招聘经理聊了一个多小时
他们有两个大项目
其中一个我三年前就在前雇主那儿作过了,并申请了专利
另一个刚刚在现公司申请了专利,并写了篇文章,已被一家需要同行评审的杂志接收发表,
我觉得像我这种背景,只要愿意去,八层得给我了
面试日程表寄来后,我曾犹豫过
因为面试是安排在午饭后一点开始
知道这边有很多STARTUP爱借面试的名义去问或偷别人的解决方案
同时吝啬得连一顿午饭都不舍得管,
觉得他们没诚意
本想临时放他们鸽子
转念一想,前一阵他们公司在对待员工等方面的排名一直很靠前
也是挺有名的大公司,又决定去了
第一个聊的是一个小箩卜头,
问了一个题目,用了一堆他那个背景习惯用的名词
开始不知道他在说啥
等弄清他的问题后,发现那题4年前就被问到过
后来被收到我给别人讲课的材料中
自然难不倒我
第二个聊的是他们的VP TECHNOLOGY, HIRING MANAGER ... 阅读全帖 |
|
b*****o 发帖数: 715 | 35 Gradient boosting和random forest这个问题我不觉得奇怪。在现实中,GB和RF就是两
个互相竞争的算法,而且形式还真有点像,都是additive,只不过一个adaptive,一个
是non-adaptive。
楼主的回答实在太跳跃了。RF有两个idea,一个是bagging,还有一个就是帖子里提到
的每个split可以做stochastic fraction的sample。我看了两遍才明白楼主是在说这个。
不过就我的愚见,其实这个idea还真不是两个算法的区别所在,因为gradient
boosting里也完全可以用这个trick的。有些版本的实现,GB和RF就都是有这个参数的。
表, |
|
m******a 发帖数: 77 | 36 接受批评!
关于 BOOST RESIDUAL 一说, 请参见前面 DATA2014 兄的回贴
遵照VICTOR前辈的建议, 在这儿多说两句, 同一篇文章, 每个人的理解可能都不一样
本人学 GRADIENT BOOSTING, 读的是这篇
http://statweb.stanford.edu/~jhf/ftp/trebst.pdf
看看他的公式 6-10 或者更简单, 他的 ALGORITHM 2
就应该很清楚他是在 BOOST RESIDUAL - 试图每一步都减少RESIDUAL
兄台既然已经写过, 想必理解必然正确, 不过可能是从另一个角度
另外, 本人理解, HMM 和 KALMAN FILTER 还是有差别的,
至少在程序算法实现和应用上, 一分立, 一连续 - 没见谁拿 KALMAN FILTER 去作 NLP
模型,
也没见谁拿 HMM 去预测股价吧 - 除非是我孤陋寡闻, 虽然都可归为态空间模型
还有, 通常说的BOOSTING, 如ADA, EPSILON, 和GRADIENT BOOSTING, 有些细微差别
也可能兄台提及的是通常的BOOSTING
通常的BO... 阅读全帖 |
|
w***g 发帖数: 5958 | 37 1. gradient越propagate越小怎么办?
2. gradient越propagate误差积累越多怎么办?
我随便问问,都是实际应用中立刻会碰到的问题。
chain |
|
n*****3 发帖数: 1584 | 38 Niu, Feng, et al. “Hogwild!: A lock-free approach to parallelizing
stochastic gradient descent.” Advances in Neural Information Processing
Systems 24 (2011): 693-701. (algorithm implemented is on p.5) https://papers
.nips.cc/paper/4390-hogwild-a-lock-free-approach-to-parallelizing-stochastic
-gradient-descent.pdf
have not really looked into the souce code yet; but it use up 20+ threads. |
|
d***e 发帖数: 193 | 39 【 以下文字转载自 JobHunting 讨论区 】
发信人: ISphoenix (beta3), 信区: JobHunting
标 题: Data scientist / Machine Learning Engineer 相关面试题
关键字: data scientist,machine learning
发信站: BBS 未名空间站 (Sun Oct 19 17:31:36 2014, 美东)
去年我找工作的时候发现板上针对data scientist,machine learning engineer面试
题总结很少,所以尽量申请了很多公司面试相关职位,想看看行业里这个方向都在问什
么。有幸去过不少地方面试,现在把那些题目整理整理(全部来自Amazon, Microsoft,
Yelp, Pinterest,
Square, Google, Glassdoor, Groupon的电面和onsite),希望能帮助在找相关工作的
同学们。
题目写的简略,请大家见谅
====================
1. Given a coin you don’t know it’s ... 阅读全帖 |
|
d***e 发帖数: 193 | 40 【 以下文字转载自 JobHunting 讨论区 】
发信人: ISphoenix (beta3), 信区: JobHunting
标 题: Data scientist / Machine Learning Engineer 相关面试题
关键字: data scientist,machine learning
发信站: BBS 未名空间站 (Sun Oct 19 17:31:36 2014, 美东)
去年我找工作的时候发现板上针对data scientist,machine learning engineer面试
题总结很少,所以尽量申请了很多公司面试相关职位,想看看行业里这个方向都在问什
么。有幸去过不少地方面试,现在把那些题目整理整理(全部来自Amazon, Microsoft,
Yelp, Pinterest,
Square, Google, Glassdoor, Groupon的电面和onsite),希望能帮助在找相关工作的
同学们。
题目写的简略,请大家见谅
====================
1. Given a coin you don’t know it’s ... 阅读全帖 |
|
c****t 发帖数: 19049 | 41 gelman的bda是大号经典没错但是出了名的难读难懂。等你从别的地方全学会了回头拿
这当reference可以。没基础的话看看jeff gill的bayesian methods吧
其实找工作刷题才是王道(不只是leetcode),进工业界干活不用懂这么多。多讨论面试
题更有用。统计版不让讨论面试题也是我开此版的主要原因之一。
另外mcmc是从理论物理来的。跟统计一根毛的关系都没有。真想学懂直接看yann lecun
有关energy based models的东东也许能明白的更快一些。我一向认为灌统计水的大谈
特谈bayesian,扯神马frequenists vs. bayesian probability,神马prior,
postorior,就是为了掩盖mcmc根本不是统计创造发明出来的事实。一般人从神马prior,
posteriror啊,bayesian probability啊学起,等你学到mcmc时早迷糊了。也就不会
多问了
mcmc是跟传统的maximum likelihood完全不同的体系。maximum likelihood说简单点就
是做优化。优化的思路就是求... 阅读全帖 |
|
c****t 发帖数: 19049 | 42 gelman的bda是大号经典没错但是出了名的难读难懂。等你从别的地方全学会了回头拿
这当reference可以。没基础的话看看jeff gill的bayesian methods吧
其实找工作刷题才是王道(不只是leetcode),进工业界干活不用懂这么多。多讨论面试
题更有用。统计版不让讨论面试题也是我开此版的主要原因之一。
另外mcmc是从理论物理来的。跟统计一根毛的关系都没有。真想学懂直接看yann lecun
有关energy based models的东东也许能明白的更快一些。我一向认为灌统计水的大谈
特谈bayesian,扯神马frequenists vs. bayesian probability,神马prior,
postorior,就是为了掩盖mcmc根本不是统计创造发明出来的事实。一般人从神马prior,
posteriror啊,bayesian probability啊学起,等你学到mcmc时早迷糊了。也就不会
多问了
mcmc是跟传统的maximum likelihood完全不同的体系。maximum likelihood说简单点就
是做优化。优化的思路就是求... 阅读全帖 |
|
c***z 发帖数: 6348 | 43 【 以下文字转载自 JobHunting 讨论区 】
发信人: ISphoenix (beta3), 信区: JobHunting
标 题: Data scientist / Machine Learning Engineer 相关面试题
关键字: data scientist,machine learning
发信站: BBS 未名空间站 (Sun Oct 19 17:31:36 2014, 美东)
去年我找工作的时候发现板上针对data scientist,machine learning engineer面试
题总结很少,所以尽量申请了很多公司面试相关职位,想看看行业里这个方向都在问什
么。有幸去过不少地方面试,现在把那些题目整理整理(全部来自Amazon, Microsoft,
Yelp, Pinterest,
Square, Google, Glassdoor, Groupon的电面和onsite),希望能帮助在找相关工作的
同学们。
题目写的简略,请大家见谅
====================
1. Given a coin you don’t know it’s ... 阅读全帖 |
|
z****8 发帖数: 13 | 44 原文链接:
http://mp.weixin.qq.com/s?__biz=MzIzODExMDE5MA==&mid=403826129&
半年前从数学专业转行到了互联网行业做数据挖掘和推荐系统,在做具体的业务的时候
遇到了一些知识点,于是自己整理出来。如果有后来人需要转行的话,可以用这份资料
来参考一下。大牛请忽视以下的内容,小白可以参考下。
从数学专业转行到工业界做数据挖掘需要的知识储备:
1. Hadoop,HIVE,SQL数据库操作。
Hive用于提取数据,做基本的数据分析。hive的基本函数,比如聚合函数,数学函数,
字符串的函数,连接表格函数等。hive的各种语句,比如if else,case等语句。
EXCEL的基本操作需要掌握,可以进行各种数据的处理、统计分析和辅助决策操作,用
熟悉了其实挺方便的。
2.编程语言
编程语言最好会python,c/c++,或者java,至少一种。做机器学习的话感觉用python
会多一些。
3.操作系统
Linux系统,脚本语言Shell。
4. 数据挖掘和机器学习的基础知识和算法
逻辑回归算法 Logistic Regression(L... 阅读全帖 |
|
z****8 发帖数: 13 | 45 原文链接:
http://mp.weixin.qq.com/s?__biz=MzIzODExMDE5MA==&mid=403826129&
半年前从数学专业转行到了互联网行业做数据挖掘和推荐系统,在做具体的业务的时候
遇到了一些知识点,于是自己整理出来。如果有后来人需要转行的话,可以用这份资料
来参考一下。大牛请忽视以下的内容,小白可以参考下。
从数学专业转行到工业界做数据挖掘需要的知识储备:
1. Hadoop,HIVE,SQL数据库操作。
Hive用于提取数据,做基本的数据分析。hive的基本函数,比如聚合函数,数学函数,
字符串的函数,连接表格函数等。hive的各种语句,比如if else,case等语句。
EXCEL的基本操作需要掌握,可以进行各种数据的处理、统计分析和辅助决策操作,用
熟悉了其实挺方便的。
2.编程语言
编程语言最好会python,c/c++,或者java,至少一种。做机器学习的话感觉用python
会多一些。
3.操作系统
Linux系统,脚本语言Shell。
4. 数据挖掘和机器学习的基础知识和算法
逻辑回归算法 Logistic Regression(L... 阅读全帖 |
|
|
m*********k 发帖数: 10521 | 47 =====================================================================
=========================Nature质疑叶诗文事件========================
============================网友回复及去信===========================
发信人: sanyanghu (忍者无敌之不去股票版), 信区: Olympics
标 题: 大家继续评论nature上的那个文章,保持热度(附我的评论)
发信站: BBS 未名空间站 (Mon Aug 6 12:03:47 2012, 美东)
大家继续跟进!!
(my recent comments):
I am happy to see that Nature started to do the right thing by adding
Laijiang's comments which was intentionally deleted before for some reasons.
N... 阅读全帖 |
|
m******v 发帖数: 66 | 48 Optimal design and preparation of silicon nitride ceramic radome material
with gradient porous structure
Fei Chen, Ling Li, Faqiang Yan, Qiang Shen, Lianmeng Zhang
INTERNATIONAL JOURNAL OF MATERIALS & PRODUCT TECHNOLOGY 卷: 39 期: 1-2
特刊: Sp. Iss. SI 页: 72-82 出版年: 2010
http://inderscience.metapress.com/app/home/contribution.asp?referrer=parent&backto=issue,6,17;journal,2,62;linkingpublicationresults,1:110879,1
http://www.inderscience.com/search/index.php?action=record&rec_id=34261&prevQue... 阅读全帖 |
|
|
z**********u 发帖数: 754 | 50 jetstream ~150mph (~60 to ~250 depending on temp gradient)
jet airliner ~550mph |
|