Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Time Series Prediction Using LSTM Deep Neural Networks

altumintelligence.com

Post reply on HN
Print view View original

31–40 of 72 posts

Re: Time Series Prediction Using LSTM Deep Neural Networks

#31

Post by jkabrg »

zwaps wrote: post #28

I find it interesting that Computer Scientists are basically rediscovering statistics. Now when predicting time series, an issue is that most model (like ARIMA, GARCH etc.) are short-memory processes. When you look at the full-series prediction of LSTMs, you observe the same thing. So in terms of Time Series, Machine Learning is currently in the mid to late 80's compared to Financial Econometrics. So if you are a CS,…

Nassim Taleb had some negative things to say about GARCH.

"GARCH does not work out of sample. It is a good story, but I was unable to use it in predicting squared deviations or mean deviations"

I haven't found it in Rob J Hyndman's forecasting tutorial either.

How does it fare in the Makridakis competitions?

Re: Time Series Prediction Using LSTM Deep Neural Networks

#32

Post by RA_Fisher »

zwaps wrote: post #28

I find it interesting that Computer Scientists are basically rediscovering statistics. Now when predicting time series, an issue is that most model (like ARIMA, GARCH etc.) are short-memory processes. When you look at the full-series prediction of LSTMs, you observe the same thing. So in terms of Time Series, Machine Learning is currently in the mid to late 80's compared to Financial Econometrics. So if you are a CS,…

It's been amazing to watch CS (really the Python community, save statsmodels and patsy) discover statistics. For a while I thought perhaps it was me and statistics that was "behind." Over time I realized that it was mostly re-invention of old ideas: one-hot encoding = dummy variables, neural networks approximating polynomial regression, etc. I decided to double-down on statistics and it's really paid off. NN / random forests and the stats-founded but CS-led approaches are very general models. That leaves statisticians a big opening because a more specific model can be chosen to obtain more accurate predictions. These days I'm positioning myself to clean-up the messes / save broken ML models. Turns out [stats] theory is very practical. :-)

Re: Time Series Prediction Using LSTM Deep Neural Networks

#33

Post by RA_Fisher »

glial wrote: post #4

Seems to me that this is almost dangerous unless the uncertainty (and therefore confidence) of the prediction can be quantified.

Yep, it is dangerous. If you're not quantifying uncertainty, you can't make safe predictions. I think this is reason for the obsession with "data cleaning" in the ML community, "outliers" aka rare observations sink general models.

Re: Time Series Prediction Using LSTM Deep Neural Networks

#35

Post by qeternity »

beagle3 wrote: post #17
jarym wrote: post #3

Why does everyone naively try to predict price? No ‘traders’ are interested in predicting it - what traders do is identify good locations to enter or exit the market. I.e. places with defined risk where you will know if you’re wrong if it goes against you by x% while you expect a y% gain if you’re right AND y>x is worth more than the number of times you’re wrong. The types of Algos that work well for this are edge id…

.. because you buy at the price, and sell at the price (spread and fees ignored for now). Which means, regardless of your philosophy, you are predicting a price change - a long signal is a prediction for positive price change; a short signal is a prediction for a negative price change. If that wasn’t true, your system would not be able to profit. Predicting price change and predicting price are semantically equivalen…

Given less than 100% certainty, traders don't want to predict price, they want to predict future distribution of price over some time period.

Source: hedge fund trader

Re: Time Series Prediction Using LSTM Deep Neural Networks

#36

Post by ahartmetz »

salty_biscuits wrote: post #25
ahartmetz wrote: post #23

Earlier quoted context omitted.

Not the same kind of experiment. An experiment in the scientific sense tweaks the process that generates the data, not the interpretation of the data. There is an inspiration / hypothesis creation step between old data and new experiment. Main differences: A hypothesis is sorta kinda like your model's coefficients, but more generally applicable. And you have no feedback loop between model coefficients and input data.…

No, it's the same. It is just about having access to control variables.

What Chomsky is saying is that the control variables don't exist until you create them because the most telling things don't happen until you have a specific hypothesis and make them happen to test the hypothesis.

Re: Time Series Prediction Using LSTM Deep Neural Networks

#37

Post by VHRanger »

jkabrg wrote: post #31
zwaps wrote: post #28

I find it interesting that Computer Scientists are basically rediscovering statistics. Now when predicting time series, an issue is that most model (like ARIMA, GARCH etc.) are short-memory processes. When you look at the full-series prediction of LSTMs, you observe the same thing. So in terms of Time Series, Machine Learning is currently in the mid to late 80's compared to Financial Econometrics. So if you are a CS,…

Nassim Taleb had some negative things to say about GARCH. "GARCH does not work out of sample. It is a good story, but I was unable to use it in predicting squared deviations or mean deviations" I haven't found it in Rob J Hyndman's forecasting tutorial either. How does it fare in the Makridakis competitions?

You shouldn't listen to N. Taleb on technical matters. He's been a classic mold crank for the last decade or so when it comes to anything serious, relegated instead to writing fluffy books on whatever he thinks is important.

Re: Time Series Prediction Using LSTM Deep Neural Networks

#38

Post by noelsusman »

beagle3 wrote: post #17
jarym wrote: post #3

Why does everyone naively try to predict price? No ‘traders’ are interested in predicting it - what traders do is identify good locations to enter or exit the market. I.e. places with defined risk where you will know if you’re wrong if it goes against you by x% while you expect a y% gain if you’re right AND y>x is worth more than the number of times you’re wrong. The types of Algos that work well for this are edge id…

.. because you buy at the price, and sell at the price (spread and fees ignored for now). Which means, regardless of your philosophy, you are predicting a price change - a long signal is a prediction for positive price change; a short signal is a prediction for a negative price change. If that wasn’t true, your system would not be able to profit. Predicting price change and predicting price are semantically equivalen…

You want to predict price, but a price prediction is useless without an accurate estimate of the error of your prediction.

Re: Time Series Prediction Using LSTM Deep Neural Networks

#39

Post by Eridrus »

RA_Fisher wrote: post #32
zwaps wrote: post #28

I find it interesting that Computer Scientists are basically rediscovering statistics. Now when predicting time series, an issue is that most model (like ARIMA, GARCH etc.) are short-memory processes. When you look at the full-series prediction of LSTMs, you observe the same thing. So in terms of Time Series, Machine Learning is currently in the mid to late 80's compared to Financial Econometrics. So if you are a CS,…

It's been amazing to watch CS (really the Python community, save statsmodels and patsy) discover statistics. For a while I thought perhaps it was me and statistics that was "behind." Over time I realized that it was mostly re-invention of old ideas: one-hot encoding = dummy variables, neural networks approximating polynomial regression, etc. I decided to double-down on statistics and it's really paid off. NN / random…

I'm curious what resources you found useful to learn stats modelling and what sorts of approaches have been useful.

On one hand, it's almost a tautoloy that specific models should be better than general models, but I worked on some 2d time series classification with a statistician and afterwards, for kicks, I replaced the entire thing with a CNN+LSTM and it worked just as well as the whole complicated model he had come up with.

Re: Time Series Prediction Using LSTM Deep Neural Networks

#40

Post by beagle3 »

beagle3 wrote: post #17
jarym wrote: post #3

Why does everyone naively try to predict price? No ‘traders’ are interested in predicting it - what traders do is identify good locations to enter or exit the market. I.e. places with defined risk where you will know if you’re wrong if it goes against you by x% while you expect a y% gain if you’re right AND y>x is worth more than the number of times you’re wrong. The types of Algos that work well for this are edge id…

.. because you buy at the price, and sell at the price (spread and fees ignored for now). Which means, regardless of your philosophy, you are predicting a price change - a long signal is a prediction for positive price change; a short signal is a prediction for a negative price change. If that wasn’t true, your system would not be able to profit. Predicting price change and predicting price are semantically equivalen…

Replying to myself instead of all three replies so far:

People, prediction is a general term. Many predictors come with accuracy estimates (and outside of finance, often prediction bounds). But even if it was only one number - if you have good prediction of the expected price change, that could be sufficient to trade as it encompasses, by definition, the sun of probability of different outcomes times their magnitude.

Either E[price] or E[log price] is a single predicted value you can successfully trade with as long as you are far from your margins, and depending of course on your utility functions.

But as I mentioned, in most fields, when you talk of a “predictor”, that’s not a single number but also accuracy estimates or even a full fledged probability distribution of future events.

Post reply on HN