由买买提看人间百态

topics

全部话题 - 话题: curves
1 2 3 4 5 6 7 8 9 10 下页 末页 (共10页)
l*****r
发帖数: 2123
1
The inverted yield curve has set off alarm bells that a recession may loom
around the corner. Historically, that's been a good bet, and recessions
almost always unleash bear markets for the Dow Jones, S&P 500 and Nasdaq.
But this yield-curve inversion, which may be short-lived, looks different.
While it raises some concern about the long-term economic outlook, the
recent inverted yield curve actually is welcome in the near term.
"We may actually have seen a rather unique occurrence: the Fed actu... 阅读全帖
g******n
发帖数: 67
2
来自主题: Quant版 - 请问到底什么是 basis curve
zero curve is built by using zero rates(only one coupon payment at maturity)
. Basis curve is defined as a comparison with swap curve。For example, USD
swap curve is based on Libor 3m, using 3m deposit, euro dollar futures and
swaps.USD Libor 6m curve is a basis curve. Although there is Libor 6m swaps
traded in the market, the most liquid ones are Libor 3m vs Libor 6m basis
swaps. Thus, theoritically, you can use 6m deposit, 6m FRAs and 3m vs 6m
basis swaps(the spread is added on the Libor 6m leg... 阅读全帖
l*****r
发帖数: 2123
3
For stock market investors, an inverted yield curve is a sign that equities
could peak before an economic recession hits. It also can be a precursor to
a bear market in stocks, where equities fall 20% or more from highs.
While an inverted yield curve has been a reliable indicator of slowing U.S.
economies, the timing of an upcoming recession has been uncertain.
Historically, prolonged "curve inversions" have usually preceded major
economic slowdowns by about a year.
What Is An Inverted Yield Cur... 阅读全帖
a**********d
发帖数: 2293
4
来自主题: Faculty版 - put away the bell curve
昨天听到的,当时以为是回到了 April Fools Day
http://www.npr.org/2012/05/03/151860154/put-away-the-bell-curve
Put Away The Bell Curve: Most Of Us Aren't 'Average'
May 3, 2012
For decades, teachers, managers and parents have assumed that the
performance of students and employees fits what's known as the bell curve —
in most activities, we expect a few people to be very good, a few people to
be very bad and most people to be average.
The bell curve powerfully shapes how we think of human performance: If lots
of ... 阅读全帖
q**i
发帖数: 174
5
来自主题: Economics版 - Confusing Curves

it's not the first time we face an inverted yield curve (70's, and
for a brief period during '93-94). (Spot) yield curve isn't really
that import to people who arb along the yield curve. What's important
is forward yield curve, and forward constant maturity yield curve.
Forward curves is mkt's implied expectation of yields in the
future, and are proven empirically quite accurate.
Fixed income is really one of the two areas where people with
profound math skills can excel. The other is derivativ
D********n
发帖数: 978
6
不是件很容易的事情,感觉。
首先,如何定义坏点?如果你知道LIBOR是一群人胡乱设的,也许你会觉得所有点都是坏
点。
其次,发现坏点之后用什么数据修补也是个问题。你在spot curve上修补没问题,但是
forward curve上看起来也许就很雷人了。
没做过类似的事情,我能想象的办法是: 算forward curve。forward curve太夸张的时
候认为出现坏点。然后interpolate forward curve来进行修补。
这只是俺大概想象的办法。未必实用。姑妄言之。

curve
f*****0
发帖数: 489
7

their
the utility (benefit) to the consumer is the same at any point on the curve
but that doesn't mean the cost is the same on the indifference curve.
so if a person is to maximize his utility subject to a cost (or minimize his
cost given a utility), he is likely to prefer certain points on the curve
than others, depending on how the cost function is constructed.
in other words, which he is indifferent utilitywise on the curve, he may
prefer certain points on the curve when you factor his cost
p****x
发帖数: 1346
8
来自主题: Statistics版 - 求助: Smooth Curve
我需要将图片中new curve 改变成像old curve那样光滑, 就是smooth new curve.
或者说把new curve变成old curve那样的, 越接近越好。
用了各种办法都不成。 请大侠帮忙推荐一个方法, 最好告诉我具体怎么做, 谢谢了!
数据如下:
Age Old New
0 0.00% 0.00%
1 0.05% 0.00%
2 0.18% 4.27%
3 0.82% 5.03%
4 2.79% 10.54%
5 6.40% 23.80%
6 11.01% 34.49%
7 16.10% 44.18%
8 21.35% 51.67%
9 26.59% 58.88%
10 31.71% 64.13%
11 36.67% 68.47%
12 41.44% 72.84%
13 46.00% 74.11%
14 50.34% 76.23%
15 54.46% 78... 阅读全帖
c*******g
发帖数: 2990
9
如果刚开始不知道学生底细,可以curve。
但是我觉得长期需要curve就是故意给自己找麻烦和难为学生。
经常需要curve某种程度也说明教的不好,或者有的老师以为出难题就显得自己水平高
,挺无聊的。
我觉得考试出一两道难点能区分出水平足够了。
我从来不curve,出难了就批卷松点儿就找回来了,关键还是要了解学生,这个作业多
少能看出点儿端倪。

it.
n*******e
发帖数: 1369
10
来自主题: Accounting版 - CPA exam is not curved. 悲剧了!
http://www.accountingweb.com/item/107728
1. A Curve on the CPA Exam?
another71.com: A common myth that circulates among CPA exam candidates is
that CPA exam scores are curved according to the performance of other
candidates taking the same section. Is a candidate's score entirely
independent of the performance of other candidates? Does the AICPA employ
any curving mechanism to keep pass rates at their historical levels of 45 to
50 percent on average per section?
AICPA: The CPA exam is not curved... 阅读全帖
r****o
发帖数: 1950
11
indifference curve是指在该曲线上的任何一点用户有着相同的utlity。
那么能否假定用户选择曲线上任何一点的概率相同呢?
举个例子。10000个用户属于同一种类型(即他们的indifference curve相同)。他们消
费apple和banana。假定indifference curves的个数有限(量化),20条。能否认为每条
indifference curve上面的用户消费概率分布均匀呢?
c*********n
发帖数: 128
12
我的starting point是还没有bootstrap的rate instrument(Libor,Future,FRA,
swap)。
我觉得如果不考虑term strucutre,把每一个rate instrument(比如3m Libor,5yr
swap)拿出来单独处理它自己的time series,这个应该就跟equity time series里面
清理数据是一样的。
但是如果考虑到rate curve是term structure,是不是可以做得更好一些?比如某个点
的存在导致bootstrap的结果不正常(在spot curve上表现为某处有个点不连续,如果
是bootstrap成forward curve,会导致局部oscillation),就说这是个坏点。
如果能把两个结合做一个cross examination,是不是就能做得更精确?
发现坏点之后的处理应该不难,直接把那个点剔除就行了吧?curve 少一两个点不是什
么大问题。

是坏
v*******e
发帖数: 506
13
来自主题: Statistics版 - 问一个用stata做roc curve 的问题
问一个用stata做roc curve 的问题, 一个multivariate logistic regression, 想做
一个roc curve, 结果用 roccomp 把所有的 自变量都做了一遍,16个curve交织在一起.
.... 其实我就只想要一个 curve.
请问应该用哪个程序??
谢谢!!
t*****8
发帖数: 157
14
来自主题: Statistics版 - ROC curve可以用来比较变量吗
今天我的同事说logistic regression中的ROC curve可以用来比较变量。比如说,
dependent variable 是y, independent variable x1 ,ROC curve是0.8;dependent
variable 还是y, independent variable 换成 x2, ROC curve是0.6. 所以x1比x2是
更strong的predictor。我印象, ROC curve只能用来看model是不是fit well, 如果
比较varible, 应该用p value. 我对还是我同事对?

发帖数: 1
15
https://www.barrons.com/articles/yield-curve-inversion-51553272263
The Yield Curve Just Inverted. What Does That Mean?
All three major U.S. stock market indexes took a downturn on Friday, as
investors responded to one of the key recession indicators: the so-called
inversion of the yield curve between the 10-year and three-month Treasuries
l****a
发帖数: 127
16
来自主题: Accounting版 - CPA exam 是如何CURVE 成绩的
有没有过来的XDJM知道CPA EXAM 是如何Curve成绩的, 是同一天或同一周的一起Curve
, 还是同一个WINDON在一起CURVE?
f***r
发帖数: 19
17
来自主题: Biology版 - Re: Real-time pcr:standard curve
1) Do your realtime primer/probe span exons? If not, genome contamination may
cause the discrepancy.
2) Narrow down your standard curve range, for example using a two or three
times dilution intead of ten.
3) It seems you use absolute standard curve method. If you just wanna compare
two samples, why not use total cDNA make standard curve and normalize to
housekeeping gene?
t**********8
发帖数: 223
18
请问slop of the dose-response curve 和 IC50 是神马关系? 分析drug
sensitivity 的时候把 slop of the dose-response curve 和 IC50 同时考虑的意义
是什么?
谢谢!
n***w
发帖数: 2405
19
来自主题: Biology版 - real-time PCR melting curve疑问?
实验室人都走了。
神通广大的生物医学版,
96孔板做RT, 一排是4个样本,1个共同基因。
这些样本最后都有threshold cycle,但是两个样本没有melting curve (none)。
这到底是为啥呢?
肯定不是cdna的问题,因为其他的primers都work的不错。
觉得也不是primer的问题,因为其他样本有melting curve (偏偏就是我的KO没有。。
。。)
google了一下没找到符合的说法,基本都是有melting curve,没有amplification。。
。。
我是不是应该跑胶check一下。。
谢谢。
s******c
发帖数: 331
20
来自主题: Biology版 - antagonist dose-response curve 求解读
schild regression plot是比较老的方法了,这两个图里做的competition binding
curves是比较现代的用来发表的东西,你为什么要转化?
就这个图来说,B很可能是negative allosteric modulator,因为competition curve
达到的plateau跟你用的agonist有关,说明B不是competitive binding。如果B是
competitive binding,理论上当B达到足够的浓度应该会完全compete掉agonists从而
最后的plateau完全是B的效果,如同A一样,agonist-independent,现在B的
competition curves plateau是agonist-dependent,说明B是non-competitive
binding。
当然,证明一个compound是allosteric ligand是复杂的,一个实验不够,往往需要多
个实验去从各个方面证明。就你贴的这个结果,很多问题,首先agonists应该不是
radioactive的,可能是fluorescen... 阅读全帖
c**a
发帖数: 94
21
来自主题: Biology版 - antagonist dose-response curve 求解读
谢谢指点,一针见血. 这两个图是用brain tissue做的, 如果用一般的细胞, 信号就不
止这么点了, 大概是4-6倍. 但是问题是,细胞做这个同样的实验, B也能把agonist
compete到最低点, 只是IC50不同. 不知道怎么解释这个现象. 我觉得这是个NAM, 但是
不知道什么才是最有说服力的实验. 你说的改变agonist dissociation实验怎么做?
还有一个问题是, 这个agonist是个endogenouse peptide, 跟已知的合成的小分子
agonist位点很可能不一样(根据文献). 这个peptide agonist 有两个问题, 一个是如
果做shift agonist curve的话(用不同浓度的PAM), 在脑组织里不好做, 因为它的
curve没有plateau, 没法算alpha值. 另一个就是它在浓度高的时候不specific, 也就
是为什么它不plateau. 所以难道我们只能用细胞证明它是PAM 吗? 还有peptide
dissociation怎么做, 要H3标记的吗?

curve
l*****0
发帖数: 299
22
原始数据点yi的std err会影响fitted curve的confidence bound吗?
每个原始数据点yi的std err SEi在做regression y(x) = fitResult(x) + ep(x) 时候
会转化成weights: wi = 1/Ni/SEi^2。

请问这些原始数据点的std errs会影响fitted curve的confidence bound吗?
知觉上好像是原始数据点的std errs越大,fitted curve的confidence band越宽,不
知是否正确?
谢谢!
f*******r
发帖数: 383
23
Can anyone here kind enough to help me find the following paper?
Please kindly send the paper to c*********[email protected].
Thanks a million.
Fitting yields curves and forward curves with maximum smoothness.
K.J. Adams and D.R. Van Deventer.
The Journal of Fixed Income, 4(1):52-62, June 1994.
j****e
发帖数: 140
24
【 以下文字转载自 Quant 讨论区 】
发信人: jejune (孑孓), 信区: Quant
标 题: given level, slope and curvature, what will be the curve?
发信站: BBS 未名空间站 (Fri Oct 26 15:59:40 2007)
i need a curve that has given level, slope and curvature at a given point and pass anthoer pont.
what would be the simplest way to construct such a curve?
PT
发帖数: 148
25
来自主题: Mathematics版 - number of inflexion points for a Bezier curve
A question puzzles me.
For a cubic curve
y(x) = ax^3 + bx^2 + cx + d
we want to find the inflexion point.
We differentiate the equation with respect to x twice to obtain
y'' = 6ax + 2b
Since at inflexion point, curvatures change signs, i.e. y'' = 0.
then
6ax + 2b = 0 ----> x = -b/(3a)
which means we have only one inflexion point for cubic curves.
However, in a website, it is said for cubic Bezier curves, there are at most
two inflexion points.
Why is that?
Thank you.
S*******r
发帖数: 11017
26
来自主题: Quant版 - 请教FWD CURVE的MODELING问题
嗯就是说,现在有个INDEX的SPOT PRICE,然后要MODEL出未来两年的FORWARD CURVE,用来
算基于这个INDEX的某TOTAL RETURN SWAP的价格.利率的FWD CURVE可以用
BOOTSTRAPPING来做(基于市面上现有的Eurodollar Futures的价格),但是这个INDEX的
FWD CURVE要怎么MODEL啊?又不是S&P IDX有FUTURES可以参考...是不是要用某个MODEL
来SIMULATE几千几万次然后取均值?
i****e
发帖数: 78
27
来自主题: Quant版 - how to build commodity forward curve?
forward curve does not exist yet, only have spot price.
I don't need to model forward curve dynamics. how
do you model forward curve?
q*x
发帖数: 62
28
use a basis curve to adjuster 3 month curve; or just use the cash rate and
swap rate to construct the 12 month curve.

future
b*********n
发帖数: 182
29
来自主题: Statistics版 - How to put confidence interval on a KM curve?
Sorry, don't have Chinese input here.
Is there anyone who knows how to put a confidence interval curve on a KM
curve? (kind of get a confidence range for all the points on a KM curve)
Thanks.
w********r
发帖数: 253
30
【 以下文字转载自 Quant 讨论区 】
发信人: whatsummer (不理猫@St Trinians), 信区: Quant
标 题: brokered CD curve and swaption curve (help help!!!!!!!!!!!!!!!)
发信站: BBS 未名空间站 (Thu Jan 26 13:46:57 2012, 美东)
The spread between brokered CD and swaption rates
1. is there mean reversion pattern?
2. how does brokered CD respond to the baseline rates?
3. what do the response lags look like under different rate levels
Is there any previous research on this? any clue on which direction/key
words to start with?
p****x
发帖数: 1346
31
来自主题: Statistics版 - 请问怎样 Smooth Curve?
我不知怎样将chart贴到这里, 所以只能把数据列到下面, 麻烦帮我的同学在excel里
看一下这个curve的形状。
我需要将这个curve变得smooth一些, 就是像一个抛物线那样光滑的, 纵轴起点是0,
终点是100. 横轴是0到42.
不知应用什么方法? 试过moving average, 不起作用。是否可用R software? 怎样用?
Age Timing Curve
0 0.00%
1 0.00%
2 4.27%
3 5.03%
4 10.54%
5 23.80%
6 34.49%
7 44.18%
8 51.67%
9 58.88%
10 64.13%
11 68.47%
12 72.84%
13 74.11%
14 76.23%
15 78.06%
16 78.61%
17 78.77%
18 79.86%
19 80.22%
20 81.50%
21 82.71%
22 83.88%
23 84.98%
24 ... 阅读全帖
l*****0
发帖数: 299
32
原始数据点yi的std err会影响fitted curve的confidence bound吗?
每个原始数据点yi的std err SEi在做regression y(x) = f(x) + ep(x) 时候会转化成
weights: wi = 1/Ni/SEi^2。

请问这些原始数据点的std errs会影响fitted curve的confidence bound吗?
知觉上好像是原始数据点的std errs越大,fitted curve的confidence band越宽,不
知是否正确?
谢谢!

发帖数: 1
33
The U.S. Yield Curve Just Inverted. That’s Huge.
https://www.bloomberg.com/opinion/articles/2018-12-03/u-s-yield-curve-just-
inverted-that-s-huge
O**********g
发帖数: 3962
34
那个flat curve就是扯谈。
第一,在武汉爆发后,在1个多月里竟然没有想到赶紧扩建医疗资源,赶紧囤积口罩防
护服呼吸机,既然已经知道病毒扩散无法阻挡,不提高医疗最大负载能力上限这真是无
法自圆其说。
第二,以目前的医疗负载能力上限,如果成功flatten curve,那就意味这这病毒要传
播好几年后才能群体免疫,这就意味着几年的lockdown,这他妈的经济大萧条得是千年
一遇了。
a**********d
发帖数: 2293
35
来自主题: Faculty版 - put away the bell curve
我觉得他们把能力和成就混为一谈了。
拿跑百米为例,如果随机选个大样本,用跑完花多少时间来衡量"performance",那么
这个分布基本上会是bell curve。但如果用拿了多少块奥运奖牌来衡量“performance"
,显然不可能是normal curve。前者衡量的是能力,后者衡量的是社会环境所界定的成
就。
Robert Frank 那本讲 winner-take-all 的书1995年就出版了,说的道理非常相似,这
次不知道是这两个心理学家哗众取宠,还是NPR这个piece给整歪了。
j**l
发帖数: 337
36
来自主题: Faculty版 - put away the bell curve
昨天我也听了这个NPR广播. 我们的学生成绩分布也不是bell curve了.double bell
curve很常见. 就是一堆人拿A, 一堆人拿F,D. 中间没几个人.
d******e
发帖数: 1424
37
来自主题: LosAngeles版 - Getting Ahead of the Curve (转载)
【 以下文字转载自 SanDiego 讨论区 】
发信人: diketene (呆客厅), 信区: SanDiego
标 题: Getting Ahead of the Curve
发信站: BBS 未名空间站 (Sun May 23 21:11:37 2010, 美东)
The 9th Annual Symposium on Biopharmaceuticals: San Diego Bio-Pharma
Conference 2010
Getting Ahead of the Curve
Date: June 5th, 2009, 8:00am – 4:45pm
Location: Hilton, San Diego/Del Mar
Website: www.sdbiopharma.org
Registration: https://www.123signup.com/event?id=mzgxr $30 for early birds(
by May 28), $50 before end of June 2, $80 for onsite registration, stud
f******n
发帖数: 198
38
不是说curve不好,是说forced curve不好,就是说你一个组里不干活儿的都已经全裁
掉了,还逼着你硬要找一个bottom 10出来。而你隔壁组里可能一堆不干活儿的,也只
要裁掉10%。
d******0
发帖数: 22800
39
来自主题: Running版 - 今天在gym里试了试curve
是啊,加个curve就卖这么贵,是很没有道义的事情,100刀平的,curve的话500刀碰顶
了。赶紧降价,我就跳。
y*b
发帖数: 3190
40
"
We saw this before from Sony and now there is a new patent from Nikon that
seems to describe the very same thing: A curved image sensor! The advantages
of such a sensor would be:
You can design more compact lenses
The Sony patent said there is a 1 stop advantage in making such lens designs
No artifacts or image quality deterioration on corners
My guess is that if that patent will ever be implemented ina real products
this will happen on a fixed lens camera.
"
http://www.mirrorlessrumors.com/al... 阅读全帖
s********a
发帖数: 1680
41
体验区别大吗 我现在是flat的34,距离1米左右(如果再远字又需要调大了)感觉有点
看不过来,比较需要扭头,curve的能明显改善吗?另外curve的会不会有时候看有些图
像会感觉失真?
c***y
发帖数: 367
42
【 以下文字转载自 Linux 讨论区 】
发信人: canny (小桔子), 信区: Linux
标 题: 请问Linux底下有没有最简易的show 2D x-y curve的工具
发信站: BBS 未名空间站 (Tue Dec 4 00:28:24 2007), 转信
运行Fluent 软件产生的 Lift vs. iteration 数据文件如下:
希望有个简易的命令能够直接show 这个 2D X-Y curve
向各位大侠请教了.
有个叫 xmgr的可以,不过我用的server 上没这东西 :(
"Lift Convergence"
"Iterations" "Cl"
1 10.03101
2 6.6097278
3 3.2647585
4 0.67387489
5 -0.62535224
6 -1.1091322
7 -0.92205094
8 -0.43060641
9 0.067002069
s****i
发帖数: 87
43
你的前两个问题,我感觉上没有准确答案,见到类似的评论都是大家自己的推测。
考试结果,我有亲身经验,应该不是自己的绝对分数。你拿到成绩报告,就会发现,自
己各部分的百分比的平均分和最终的成绩不一样。而且列示出来的标准成绩(应该是大
家的平均分)的平均分也都各科高低不同,比如审计一般都在80多分了。所以我推测,
因该是有很多权数来计算我们的最终分数,题目的难度,参考的人群,因该还有其他比
较因素,都是有影响的。否则,当场你的成绩就出来了,就和GMAT一样了。大多人都相
信是curve,但是curve的方法不确定。
n*******e
发帖数: 1369
44
来自主题: Accounting版 - CPA exam is not curved. 悲剧了!
So even they say not curved, it is still curved?
s****i
发帖数: 87
45
来自主题: Accounting版 - CPA exam is not curved. 悲剧了!
成绩报告是告诉考生的各部分正确的百分比和达标的百分比。四门成绩的达标平均分都
不同,相差不小的。考生自己的百分比和最后的成绩也是有差距的。
还有一个证据能说明成绩不是绝对分数:最明显的就是会计,有相当的人最后的大题只
做了一道或者两道都乱选的,最后居然能拿到90左右的高分,你就可判断这不是绝对成
绩了。
综上两点,所以很多人推测成绩是curve 出来的,但方法应该比较复杂,不是简单的用
参加考试的人来curve。网站上的话就是外交辞令。
h******u
发帖数: 602
46
实验室送出一文章,REVIEWER要求显示某一real-time RT-PCR的amplification curve
。我们 用的是ABI7000机子,不知如何能把curve直接COPY下来?打印出来再扫描太麻
烦了,质量也差了。。。
s*****r
发帖数: 59
47
来自主题: Business版 - flat yield curve
given a yield curve, how to calculate a flat single yield from it ( a flat
yield curve). is it just geometric mean? if not, how? Thanks.
l******n
发帖数: 1204
48
来自主题: Economics版 - 请教: labor supply curve
1,想问问那些因素影响labor supply curve的弹性?
2,直觉上,中国的labor supply curve弹性小,比较陡峭,是否正确?
3,如果弹性小,加税,tax incidence会更多的有个人负担。
如果弹性小,缩减个人所得税,个人得到的实惠应该比较大。
结论,如果减税,可以有效提高个人收入。
这个对吗?
谢谢。
s*****g
发帖数: 352
49
来自主题: EE版 - Re: Circle Curve-Fitting

Pick any two points on a circle, and you get a secant line.
Draw a straight line perpendicular to this secant line, and
it goes through the center of the circle. Now pick two other
points and repeat the above two steps. Where the two perpendicular
lines cross is the center of the circle. From that you'll know
the radius, etc etc.
Another way is to assume a center (x,y,z) and pick 5 or 6 points
on the curve, hence you get 5 or 6 radii. And your goal, for curve
fitting, is to minimize the differe
1 2 3 4 5 6 7 8 9 10 下页 末页 (共10页)