375 reputation
112
bio website
location
age
visits member for 1 year, 5 months
seen 19 hours ago

Aug
25
comment Art market specificities
I am very interesed in the data colletcor. What langage do you use ?
Aug
25
revised Art market specificities
added 77 characters in body
Aug
24
asked Art market specificities
Aug
24
comment How does this follow from the separating hyperplane theorem?
I can't manage to find one (I would have linked to it if i could). W is a subspace of $R^K$ with the first coordinate set to 0 and the k-1 others free, so it is a subspace of dimention k-1. Imagine W as a plane touching in only one point some kind of a sphere. Theyre are not stricly separated here, meaning that we accept $x.e<=0$ or $x.e>=0$ for each sets. The touching point has to be on the separation (only way to satisfy bot equations). Then the orientation: if we choose another plane than W they would cross, the only way to separate the two object is to take W as the border.
Aug
24
comment How does this follow from the separating hyperplane theorem?
well this is the particular case where the separation plane between two set is one of those set (W) so the orthogonal of the separation plane coincide with the orthogonal of one of the sets ($W^T$). The product will be positive with the use of the theorem. The formulation of the theorem is more complex than "there exist a plane that separate the two sets." In fact in the proof we build a particular vector e such that e.x > 0 for every x in the first set and e.x<0 for every x in the second set, the separation plane is then defined as the boundary (ie such that e.x = 0).
Aug
24
answered How to develop your own interest rate model?
Aug
24
answered How does this follow from the separating hyperplane theorem?
Aug
19
accepted Machine learning for non optimal behaviour
Aug
16
comment Risk-Free Rate determinant in CAPM
Could you provide more information on how they use these risk free rate ? My guess is that they choose the risk free rate accordingly to the time-step used in the considered method.
Aug
14
comment Machine learning for non optimal behaviour
> If you can find a function that replicates well enough the past exercise decisions... Well that is the point of machine learning...
Aug
12
comment Machine learning for non optimal behaviour
Basically: the output take positive values some time after there is a drop on client rates. The lag between the drop and thechange in output depends on the size of the drop.
Aug
12
comment Machine learning for non optimal behaviour
Thanks for your answer, maybe a bit broad. (see my edit). I will look at your last link a bit deeper tomorrow as DNN may be sufficient for what I wanted to do. Would you say that DNN can account for (non-linear ?) lags in the market ? Or is this 'just' a NN on a moving windows ?
Aug
12
revised Machine learning for non optimal behaviour
added 550 characters in body
Aug
8
asked Machine learning for non optimal behaviour
Jul
24
comment Should I analyze the tick data day by day?
What is your goal ?
Jul
15
comment Max Likelihood via Marquardt Optimisation
Do you have documentation for eviews ?
Jul
15
accepted Forward rates formulae
Jul
15
comment Getting the next price of a GBM with reversion
So you can use whatever you want... Typically: annual vol = 0.15, annual drift =0.05, price equilibrium price on the same level (110 and 100 for exemple). Or if you build an interface you can let them be specified by the user.
Jul
14
comment Getting the next price of a GBM with reversion
This is the difficulty... the process called calibration. It is abit difficult to get parameters from your data. Luckily there are probably people on the internet who have done it. The source of these data will depends on the asset you want to modelize.
Jul
13
comment How to get permanently growing chart within PCA
1) There is some dangerous approximations on the article you linked. 2) They lead to approximations from you: no there is no garanteed portfolio wich is almost increasing. 3) Your plot is not available anymore, but it look like you ploted the weigth of an asset inside the PCA. You should add your code if you want us to point out what is wrong.