# Tag Info

70

There seems to be a basic fallacy that someone can come along and learn some machine learning or AI algorithms, set them up as a black box, hit go, and sit back while they retire. My advice to you: Learn statistics and machine learning first, then worry about how to apply them to a given problem. There is no free lunch here. Data analysis is hard work. ...

28

My Advice to You: There are several Machine Learning/Artificial Intelligence (ML/AI) branches out there: http://www-formal.stanford.edu/jmc/whatisai/node2.html I have only tried genetic programming and some neural networks, and I personally think that the "learning from experience" branch seems to have the most potential. GP/GA and neural nets seem to be ...

19

Because of: The (extreme) dominance of noise over signal The prevalence of non-repeating patterns (many of which we know are not going to repeat) A pathetic sample size for cross-validation Regime changes due to exogenous events. These are typically in the cross-val window which makes it even worse. (GFC, financial integration, trade law changes, interest ...

18

By "cryptography" you mean information theory. Information theory is useful for portfolio optimization and for optimally allocating capital between trading strategies (a problem which is not well addressed by other theoretical frameworks.) See: J. L. Kelly, Jr., "A New Interpretation of Information Rate," Bell System Technical Journal, Vol. 35, July ...

17

First, let's speak about perceptrons in general: their input $X_0$ is a $K$-dimensional vector. So if you want to use $(P_{bid}(t),P_{ask}(t), Q_{bid}(t),Q_{ask}(t))$, it would mean that without any effort (but later we will see that is would be better to do some efforts, as usual): $$X_0(t)=(P_{bid}(t),P_{ask}(t), Q_{bid}(t),Q_{ask}(t))'\in\mathbb{R}^4$$ ...

15

I would say in the context of trading in general (for HFT see my comment above) further developments of recurrent neural networks (RNN), e.g. so called historical consistent neural networks (HCNN) together with forecasting ensembles, are state of the art. I published an article on that which will be published this month by Springer Verlag (Zimmermann, ...

13

Stanford University has a free online course in machine learning with video lectures, problem sets, and even a promise of online help with coursework from Stanford faculty. This course provides a broad introduction to machine learning, datamining, and statistical pattern recognition. Topics include: (i) Supervised learning (parametric/non-parametric ...

12

This is a very interesting question. I believe it is getting a lot of up-votes from people who have wondered the same thing and don't know where to begin, whereas you have at least laid out a reasonable-sounding plan. I commend you for that. However, it is not clear to me what you're trying to learn by posting this question. In my opinion, the plan you ...

11

Two aspects of statistical learning are useful for trading 1. First the ones mentioned earlier: some statistical methods focused on working on live datasets. It means that you know you are observing only a sample of data and you want to extrapolate. You thus have to deal with in sample and out of sample issues, overfitting and so on... From this viewpoint, ...

11

The short and brutal answer is: you don't. First, because ML and Statistics is not something you can command well in one or two years. My recommended time horizon to learn anything non-trivial is 10 years. ML not a recipe to make money, but just another ens to observe reality. Second, because any good statistician knows that understanding the data and the ...

11

In a very, very general sense, what Renaissance Technologies does well [and others try to do, many do less well] is understand where the "true" signal is (i.e. where prices should be) and what is noise (i.e. over-/under-reactions by others in the market) in the total signal of market prices. Generally, trading profits are made by taking the opposing ...

10

Most contemporary NN systems are just made to use the raw price time series for input (maybe with some kind of simple normalization), but for my thesis I wrote a system which traded equities with an ANN with technical indicator inputs (MAs, MACD, even pattern matching for stuff like Head-Shoulders, support levels, etc.). So at least conceptually it's ...

10

As the others have already mentioned, this is a very broad question. Anyway, as a starting point there are some blogs that come to my mind that have some up to date high quality content on these issues from time to time: http://quantivity.wordpress.com/ http://epchan.blogspot.com/ http://www.automated-trading-system.com/ ...

10

I think R's CRAN Task Views on Machine Learning is an excellent resource for beginners moving to advanced algorithm traders. It is well-structured, broad, up-to-date, and ready-to-use! http://cran.r-project.org/web/views/MachineLearning.html I believe all advanced quantitative traders already know this. But I haven't seen anyone post it here and Flake's ...

9

I honestly think that most people who could be able to answer to this question simply won't either because they actually work for Renaissance, or because they work in a top quant hedge fund and they'll keep it a secret. I discussed this topic once during an interview and the guy said "we'll discuss this further if you get the job" lol. About papers, I'm ...

8

The predictor variables would consist of the input layer to the neural network. The output layer would consist of your target. You need to specify the hidden layer, number of nodes per layer, the learning algorithm, and the learning algorithm stopping criteria. Typically inputs are normalized (first-differenced, z-scored, etc.) before inputting into the ...

8

If you're looking for Java or C/C++/C#, then you will have a much harder time with this than if you looked at R, Matlab, or Python (with Scipy). For those other languages, I recommend: Java: Weka is one of the most complete data mining libraries out there. Fortunately, it also comes with a very good book -- "Data Mining: Practical Machine Learning ...

8

You should consider an unsupervised learning algorithm such as K-nearest neighbor ('KNN'). KNN will measure the distance amongst the observations in your space. You can and probably should consider alternative distance functions (besides euclidean) particularly if you are clustering on features such as returns which have outliers. There are quite a few ...

6

A good resource for open-source statistical learning / machine learning libraries is mloss.org.

6

If you are programming in C#, you may have a look at AForge.NET and Accord.NET too

6

An interesting pick if you'd like to use Python within the Numpy/Scipy environment is scikits.learn. And an other viable Java package is Apache's Mahout.

6

From this site's perspective, I think nothing would be better than a ML.SE. Finally, we got one awhile ago. UPDATE: Unfortunately, Machine Learning is merging into Cross Validated. To learn more detail, click here." I have no idea why SE admin was rush to merge ML into CrossValidated. Not a fan of it (Orz). I personally prefer a separate site. FYI, ...

6

One relevant paper is: Shenoy, C. and Shenoy, P.P., Bayesian network models of portfolio risk and return, 1999. PDF

6

First, we are few quants and academics to use the full toolkit of machine learning: stochastic algorithms, to optimal trading. Here are at least two papers: Optimal split of orders across liquidity pools: a stochastic algorithm approach, Sophie Laruelle (PMA), Charles-Albert Lehalle, Gilles Pagès (PMA) Optimal posting distance of limit orders: a stochastic ...

5

One basic application is predicting financial distress. Get a bunch of data with some companies that have defaulted, and others that haven't, with a variety of financial information and ratios. Use a machine learning method such as SVM to see if you can predict which companies will default and which will not. Use that SVM in the future to short ...

5

I will break up your question in to some parts to make answering easier. "people use various economic indicators with their networks (moving average, MACD, etc.) However, how do these come into play in a NN context?"--the 'indicators' MA, MACD etc. come from the data. They are measures of the data capturing some aspect. You could try to capture/replicate ...

5

I have been learning more about speech recognition motivated by its application to financial forecasting. I have identified a couple connect points. Turns out each of these tools can and are regularly used in financial modeling as well. Use of Markov Models Use of Fourier transforms (sine/cosine decompositions) Use of component analysis

5

Quant Guy's answer is quite informative for your question already. Just to add few other things: instead of figuring out the choice of features by your own brain, you could also use machine learning techniques to help in extracting the 'features' for your specific purpose, e.g. risk modeling or returns forecasting or portfolio construction as mentioned by ...

4

The Brown et al. paper and its connection with trading is discussed here: http://bit.ly/kXyF99

Only top voted, non community-wiki answers of a minimum length are eligible