都来了，测测你有多阴暗！

[duality]

y: 你发的负面帖子的数量 （负面帖子指尖酸刻薄，谩骂攻击，讽刺挖苦等非建设性的帖子）

x: 你发的正面帖子的数量 （正面帖子指友好的，建设性的，善意的帖子）

你心里阴暗的可能性是：4^y/(4^x+4^y)

注：4^x 的意思是４的x 次方。

计算出来的数=１，对应与心里阴暗的可能性百分之百；
计算出来的数=０，对应与心里阴暗的可能性是０。

数学模型见
http://bbs.comefromchina.com/forum5/thread809337.html

 

[ZhiZuChangLe]

中华文化之精髓就在于“阴阳”二字，许多西人都知道这俩字。可怜LZ貌似中国人，却不知之。

 

[duality]

中华文化之精髓就在于“阴阳”二字，许多西人都知道这俩字。可怜LZ貌似中国人，却不知之。

y:=y+1 for you?

 

[ZhiZuChangLe]

y:=y+1 for you?

随您怎么说，俺不在乎。俺只觉得您很可怜。

 

[manifold]

y: 你发的负面帖子的数量 （负面帖子指尖酸刻薄，谩骂攻击，讽刺挖苦等非建设性的帖子）=500+1%*500=505

x: 你发的正面帖子的数量 （正面帖子指友好的，建设性的，善意的帖子）=500

你心里阴暗的可能性是：
4^y/(4^x+4^y)=4^505/(4^500+4^505)=0.00097561


For people who had never posted anything,
你心里阴暗的可能性是 ........ 50%

 

[duality]

y: 你发的负面帖子的数量 （负面帖子指尖酸刻薄，谩骂攻击，讽刺挖苦等非建设性的帖子）=500+1%*500=505

x: 你发的正面帖子的数量 （正面帖子指友好的，建设性的，善意的帖子）=500

你心里阴暗的可能性是：
4^y/(4^x+4^y)=4^505/(4^500+4^505)=0.00097561

you calculated wrong. It should be close to 1.

 

[manifold]

No offence, just interested in math and willing to discuss.

4 is an interesting constant in your model.

4^y/(4^x+4^y) = 1/(1+4^(x-y))

suppose x-y<0

what if the constant is less than 1 instead of 4?

Go back to the way you build the model:

每个心理阴暗的人每次发帖以概率 a 发负帖
每个心里亮堂的人每次发帖以概率 b 发负帖

I guess 4 comes from a or b.

As being a 心理阴暗的人, he will set "a" to be ...?
Similar to 心里亮堂的人.

I only know skin and fur of probability and I think Bayes' theorem is very difficult. This theorem is sort of inversing the order of what happened in the real world. It is applied in the area of epedimiology and it is reasonable to apply Bayes' theorem here (am I right?), but how?

 

[manifold]

you calculated wrong. It should be close to 1.

EDU>> 4^500/(4^500+4^505)

ans =

9.7561e-004

by Matlab. Let me try other numbers

 

[manifold]

 

[manifold]

oh ... I see. I mixed x and y. haha.

What I want to say is, a person who only needs 1% more of 发的正面帖子的数量 than 发的负面帖子的数量, and 帖子的数量 is large enough,
then he is likely a 心里亮堂的人 (see above figure). However, if the constant is than 1, he is still a 心理阴暗的人. WLOG, y>x and constant is less than 1, .....

 

[duality]

oh ... I see. I mixed x and y. haha.

What I want to say is, a person who only needs 1% more of 发的正面帖子的数量 than 发的负面帖子的数量, and 帖子的数量 is large enough,
then he is likely a 心里亮堂的人 (see above figure). However, if the constant is than 1, he is still a 心理阴暗的人. WLOG, y>x and constant is less than 1, .....

For your interest, I posted the derivation of the formula at
http://bbs.comefromchina.com/forum5/thread809337.html

In case that there are equal probabilities at which a person is ``dark-minded'' and ''bright-minded', the constant (4) comes only from the ratio a/b. So as long as the person have y>x, then he will have a higher chance to be dark-minded. This comes true in general as long as the a/b>1. It does not make sense to make a/b<1 in my opinion, since it would correspond to the model where dark-minded people are more likely to publish postive postings than negative postings.

Also under this model (with reduced form of equation), the absolute value of (y-x) decides how much the inferred probabilty deviates from 0.5. y=10 and x=5 would give the same result as y=505 and X=500.

After seeing your post I realized that this part of the model may not be perfect. This imperfection comes from that fact that I consider the population divides into only two classes: dark-minded and bright-minded. A better model will be to include a third class (the neutral-minded class). This should better address the issue. -- Thanks for the interest.

 

[duality]

BTW, I like your user name. A demonstration of good math background.

 

[说不说]

数学只是个抽象概念，用中文来说应该属于‘术数’概念，中国古代就很发达了，只可惜很多东西失传了。
LZ用数学来推断‘人性’，实属一大创意... 可敬啊~

只是末学愚见，心性浩瀚，非数学所能概... 而人之为善为恶，观其言、行、达者自明，何须用公式来推断。

网上对于某些人来说，只是虚拟的世界，所以便‘为平常所不为，言平素所不言’，不畏其因果，可惜可叹....
LZ的一番苦心只怕是无人领情呢。

 

[duality]

数学只是个抽象概念，用中文来说应该属于‘术数’概念，中国古代就很发达了，只可惜很多东西失传了。
LZ用数学来推断‘人性’，实属一大创意... 可敬啊~

只是末学愚见，心性浩瀚，非数学所能概... 而人之为善为恶，观其言、行、达者自明，何须用公式来推断。

网上对于某些人来说，只是虚拟的世界，所以便‘为平常所不为，言平素所不言’，不畏其因果，可惜可叹....
LZ的一番苦心只怕是无人领情呢。

所言极是。当今数学虽然非昔日可比，人心百态，哪里可以几行公式推断。弄个公式玩玩，无非是漫画一下某些人的丑态而已。

 

[manifold]

数学只是个抽象概念，用中文来说应该属于‘术数’概念，中国古代就很发达了，只可惜很多东西失传了。
LZ用数学来推断‘人性’，实属一大创意... 可敬啊~

只是末学愚见，心性浩瀚，非数学所能概... 而人之为善为恶，观其言、行、达者自明，何须用公式来推断。

True. But applying math on 人性 makes a lot of fun. I had seen a math model which involved ordinary differential equations to describe "love affair".

I am a bit sick of reading arguments about FLG, Hu, ....
Seeing a math model at CFC is like seeing an oasis.

By the way, I replied several posts in this thread and tried to explore the equation because I felt this equation is so familiar to me. While I was driving, I then suddenly remembered that I saw this equation few years ago. It is called Hill Equation and here is the link showing it's properties

http://en.wikipedia.org/wiki/Hill_equation