9

The bias comes from the paper Stambaugh (1999) and has nothing to do with small sample bias. It has to do with point (1) below. The argument goes as follows: Typical lagged explanatory variables for stock-return regressions are correlated with contemporaneous stock returns This contemporaneous correlation biases forecasting regressions First review OLS ...


8

Let me start with a simple example. Suppose you have a dividend strip that pays an unknown dividend $D_T$. The gross return (something like 1.05 and NOT 5%!) on this security is, by definition, $$R_{t\to T} = \frac{D_T}{P_t}$$ where $P_t$ is the current price of this security. If we use lowercase letters to denote logs (i.e., $\log D_T = d_T$ etc..) we can ...


7

Without seeing your trading desk's P&L it's impossible to say whether it is predictable or not. But here are a few thoughts - There's no reason to think that it isn't predictable. In general, financial time series are hardest to predict when the represent the return stream of an investible asset. A trading desk's P&L isn't really investible, so ...


6

My favorite tool is Sornette's own Finanical Crisis Observatory: http://tasmania.ethz.ch/pubfco/fco.html If you are interested, I have developed my own tool in Java and JavaCL which can be found here: https://thebubbleindex.codeplex.com/ Update: Code moved to github: https://github.com/thebubbleindex/thebubbleindex


5

Sorry, but despite being used as a popular example in machine learning, no one has ever achieved a stock market prediction. It does not work for several reasons (check random walk by Fama and quite a bit of others, rational decision making fallacy, wrong assumptions ...), but the most compelling one is that if it would work, someone would be able to become ...


5

The mean could be the long run variance which is sig2 = fit.Constant/(1-fit.GARCH{1}-fit.ARCH{1}); I hope this explains. If not, note I ran this model through Matlab, I get different values. you can paste your m1 and m2 values and some other intermediate results so I can see why Matlab differs. EDIT: The question refers to forecasting the returns. ...


5

The graph you attached suggests that you were trying to find swings between major highs and lows. This can be done by simply finding local extrema in the price series. The concept is: find local extrema: minima in Low prices, maxima in High prices; find local extrema in the results, if swings are too short; repeat #2 until satisfied with the results. This ...


5

I have written an entire paper on this approach at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2828744 As to your specifics 1) "Volatility" as defined by variance does not exist, which is why it is changing. The first moment is undefined so the second cannot exist. See the paper as to why. Your fitted pdf will treat the outcomes as having a ...


5

There is large literature on MIDAS (mixed-frequency data sampling) models, the leading scholars being Eric Ghysels and Rossen Valkanov — google their research for references. However, the motivation for these models has mostly been to forecast low-frequency stuff with high-frequency variables, updating, say, quarterly GDP predictions as weekly ...


4

From what I have read, there are 3 popular algorithms for financial time series. Random Forests and SVMs, then followed by Neural Network Architectures. There are a couple of good papers, to name a few: Empirical Asset Pricing via Machine Learning Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500 An ...


4

Recent research A recent article by Frank Zhao is interesting to get started: Natural Language Processing - Part I: Primer. You will find more papers on this repo (too long to copy all here): nlp_papers Applications If you are looking for possible applications of current SoTA research to financial markets, here is a quick list: Equity Predict the ...


4

The correlation matrix is a very important part of modeling stock returns. It is often better to build a model that takes in multiple assets features so that it can use this correlation to its advantage. A good example of this is a VAR model from econometric. A great example in the machine learning context is the paper titled Empirical Asset Pricing via ...


4

Another way of staying "time-varying risk-premium", is saying that the risk-premium is predictable. However, that the fact that the risk-premium is predictable does not means that you can make money out of this. The best two references to understand this are: Cochrane (2008) - The dog that did not bark Goyal and Welch (2007) The first tells you what ...


4

The point of confusion may be in thinking that a predictable price process is synonymous with a mean-reverting process while using the definitions in these papers, it's actually the opposite! In the context of these papers, a random walk would be 100% predictable: the unpredictable component of a random walk (i.e. the period specific shock which has finite ...


4

There are a few exclusions that I have commonly seen: Excluding thinly traded stocks. The price that shows up in your data feed may not relate to actual tradable prices. Filtering for ADR/Pink locals. You can find stocks listed in multiple places in ways that would lead you to think that they are great for pairs trades when actually they are the same ...


3

The TA_lib Technical Analysis library here has open source code for numerous indicators.


3

The renowned CXO Advisory Group has a section "What Works Best?". Here some general information is given and many links to their research articles which e.g. summarize lots of current academic research (although most of the linked articles are behind a paywall the links to the original papers are normally provided). The article closes with "In summary, ...


3

I am not sure I perfectly understand your question, the concept of "time series with varying density over time" is not very clear. One thing is for sure, the optimal way to "feed" a neural network is a function of the type of NNet itself and of the learning method you have chosen. For time series either you believe your data are iid vectors, and you can ...


3

The accuracy of a model is only 1 factor in determining usefulness. Aside from the accuracy, it would help to determine how you would implement it in a simulated trading environment and look into the performance further. Aside from a hit ratio or accuracy, you should compute other metrics such as: The risk-reward ratio that your model realizes (not the ...


2

You can try this course on Udactiy https://www.udacity.com/course/machine-learning-for-trading--ud501


2

Blair Hull as an idea: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2609814 He says he sold his automated trading firm to Goldman for 300 million $.


2

A prediction model that is correct $50\%$ of the time can be profitable if the model gains more when it is right than it loses when it is wrong. You could simplify it like this: A trading strategy is profitable if your trades have positive expected value. Now suppose that your gains when your model is right equals the losses when your model is wrong. If ...


2

An interesting variant from Reuters (you can do your own "simulations"): https://www.breakingviews.com/considered-view/numbers-add-up-to-germany-retaining-world-cup/ Another paper from a renowned source: Probabilistic forecasts for the 2018 FIFA World Cup based on the bookmaker consensus model Achim Zeileis (achim.zeileis@r-project.org), Christoph Leitner ...


2

The Technical Analysis of Financial markets is considered as a milestone of the matter. I suggest to read that before starting to test your strategy. It explains well the use of each indicator, providing the economic reason behind that and when it is useful to use that; moreover, the book deals the stock market with mainly, as you need for. In my humble ...


2

A very good reference can be found here: http://www.asiapacfinance.com/trading-strategies/technicalindicators


2

I cannot seem to find that article for free, so here is a more generalized answer. 1.what are the hidden states and what are the observation states. The hidden states are said to be that of an unobserved parameter process following the Markov property. The observation states are generated by the hidden parameter process. The parameter process changes or ...


2

Which strategy will work better is an empirical question that depends on the data at hand. That is, you cannot prove theoretically that one approach is better than the other without some extra assumptions. I could of course back test different filtering strategies based on these two, but the more complicated, the more risk of overfitting IMO As long as ...


2

I think you have a small misunderstanding in terms of what the folks with all of the various incarnations of quantitative degrees are doing. There are always people trying to punt on the direction of the stock market for sure - and there is always an appetite with investors to try to outperform the market. What you are missing is that humanity is net long ...


2

The issue is how do you evaluate the success of your trades. If the P&L in your simulation is measured as C(t+1)-C(t) then your simulation is not completely realistic, because in real life by the time you compute the signal based on C(t), even if you do so quickly, the price C(t) will not be the current price and you will not be able to buy at that price....


2

You didn't say what you were going to do with the result and so there are two answers. First, if you do not care about the future and the only concern is to precisely match the data that happened, then you can always fit any curve by constructing a polynomial whose degree is equal to the number of data points. It will fit perfectly. It will be useless ...


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