10. Regularized Pricing and Risk Models
Вставка
- Опубліковано 29 сер 2024
- MIT 18.S096 Topics in Mathematics with Applications in Finance, Fall 2013
View the complete course: ocw.mit.edu/18-...
Instructor: Ivan Masyukov
This is a guest lecture on regularized pricing and risk models, featuring explanations of bonds, swaps, and yield curve models.
License: Creative Commons BY-NC-SA
More information at ocw.mit.edu/terms
More courses at ocw.mit.edu
Timestamps:
0:02:17 Bonds
0:17:00 Swaps
0:22:00 Yield Curves
0:42:23 Principal Component Analysis (PCA)
0:50:52 Regularized Risk Model
0:59:15 Parametric Volatility Surface
1:09:23 Regularization
This was one of the more useful lectures in this course, even though it requires quite a bit of background knowledge to follow.
Hey, did you find relevant lectures and resources for risk modelling. Would you be fine to share the same? It's just that I found your comments for Kenneth's lecture, and this one, maybe you too are finding resources for practical applications.
So far, not really, and Ken didn't come through with his promise either.
The content of this talk is really dense, not to mention that topics about fixed income are in general way harder than those about equities. There are probably enough materials for fixed-income itself to be taught in a separate class lol
This guy literally came in and winged everything. He has no idea what to follow up.
56:22 Volatility Surface
0:32:49 интересно почему спред в экспоненте? Если имеется ввиду спред между покупкой и продажей, то он же не зависит от времени
very difficult to follow...
Lost starting regularized risk model at 51:00
On 59:54 the dude said they have no memory to store 28K by 28K matrix. Assuming in float, it just took a bit over 3GB of mem. Any 16GB gaming rig can do that. Am I missing something?
Probably in order to model 28Kx28K param, you will need not only to store result's as a float but as well some temporary numbers, used in the model for each point, like PCA components, or some derivative values. So according to your calculation, if it's needed to store at list 5 float for each point it's already 3GBx5 = 15 GB, But when you make MC simulation, you need approximately 10k observation (simulated) for each point). maybe that the point, but it's very difficult to understand the speaker (even he speak Runglish, and for me as Russian it should be easier to understand him :) )
In fact :
- It was done in 2013
- And you still have to manipulate those matrices; sometimes lot of times during the day :)
44:46 I guess we are finally at that point in the future
Extremely difficult to understand and follow.
The ideas are there in the slides, that's clear. But the speaker is simply atrocious. Extremely difficult to understand and follow.
You are boss
Anyone interested in working through the course together?
I'm interested. Let's get in touch.
+
@@user-sk5fh9qo3v it's difficult to coordinate. Let me know if you get this message.
lol @ the useful links
хахах, ну если честно википедия не совсем плохой источник на математические темы
Need English language speakers to deliver a lecture in English. Listening to this video is frustrating.
Ridiculous camera work. Why does he show the freakin face of the lecturer? It's not a rhetorics class. I need to follow the slides.
Simply impossible to follow the lecture since the camera shows 1 seconds of the slide and drifts away
Suggestion: Use half of the display to show the video, and the other half to show the pdf slide.