Каменный век
Jul. 19th, 2012 07:58 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
ny_quantНе справились с матрицей со стороной в 13,000. Придется покупать новый комп.
Posted via m.livejournal.com.
ny_quantny_quant
Profile
ny_quant
Page Summary
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
Style Credit
- Style: Neutral Good for Practicality by timeasmymeasure
Expand Cut Tags
No cut tags
no subject
Date: 2012-07-19 12:30 pm (UTC)no subject
Date: 2012-07-19 03:30 pm (UTC)no subject
Date: 2012-07-19 02:07 pm (UTC)no subject
Date: 2012-07-19 03:29 pm (UTC)no subject
Date: 2012-07-19 07:59 pm (UTC)no subject
Date: 2012-07-19 04:06 pm (UTC)no subject
Date: 2012-07-19 06:00 pm (UTC)Нам сказали, что тем, кто это освоит в жизни уже ничего не страшно. Так, в общем, и вышло.
no subject
Date: 2012-07-19 06:04 pm (UTC)no subject
Date: 2012-07-19 06:12 pm (UTC)Это пять!
no subject
Date: 2012-07-19 08:51 pm (UTC)no subject
Date: 2012-07-20 01:14 am (UTC)no subject
Date: 2012-07-19 08:32 pm (UTC)no subject
Date: 2012-07-19 09:39 pm (UTC)no subject
Date: 2012-07-20 12:30 am (UTC)Дата майнингом занимаетесь?
no subject
Date: 2012-07-20 01:13 am (UTC)2. Нет. Это для Монте-Карло.
no subject
Date: 2012-07-20 01:38 am (UTC)no subject
Date: 2012-07-20 02:04 pm (UTC)For valuation, you can replace the original equation, say,
dF = CdW
where C is the sqrt of covariance matrix and bold are for vectors) with
dF = s1*P1*dw1 + s2*P2*dw2 + s3*P3*dw3
where P_i's are principal components, s_i's are their respective vols and w_i's are scalar Weiners. That way you need to generate order(s) of magnitude fewer randoms and you'll get faster convergence.