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

Using a Keras Long Short-Term Memory Model to Predict Stock Prices

heartbeat.fritz.ai

Post reply on HN
Print view View original

21–30 of 178 posts

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#21

Post by this_user »

AznHisoka wrote: post #13
joshgel wrote: post #12

I wonder if someone could comment on how AI is used at big investment banks? I assume its much more than technical analysis (like this model). I imagine that NLP might be helpful for quickly ingesting a news feed and then making buy/sell decisions based on the news in less time than humans can read and react?

As you mentioned, one way is to quickly ingest earnings announcements from press wires/company websites. As soon as the earnings report hits the wire, they want to immediately known how big a miss/beat was it, and whether they should immediately dump it or buy more. Other examples include analyzing the words/sentiments in earning calls, to see if certain words indicate a bearish or bullish signal. For example if a CE…

I don't think you need AI for that, though. There are news feeds that are specifically designed to be machine readable. All you need to do is parse it and compare the actual numbers to the expected ones, then calculate the delta for your pricing model. All of this is pretty simple. The real challenge is being faster with this than the competition.

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#22

Post by AznHisoka »

this_user wrote: post #21
AznHisoka wrote: post #13

Earlier quoted context omitted.

As you mentioned, one way is to quickly ingest earnings announcements from press wires/company websites. As soon as the earnings report hits the wire, they want to immediately known how big a miss/beat was it, and whether they should immediately dump it or buy more. Other examples include analyzing the words/sentiments in earning calls, to see if certain words indicate a bearish or bullish signal. For example if a CE…

I don't think you need AI for that, though. There are news feeds that are specifically designed to be machine readable. All you need to do is parse it and compare the actual numbers to the expected ones, then calculate the delta for your pricing model. All of this is pretty simple. The real challenge is being faster with this than the competition.

You're absolutely right. None of this is AI. Wall Street is notorious for making stuff sound more impressive then it really is.

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#23

Post by ElBarto »

Isn't the concensus these days that every bit of information is squeezed out so that the stock price always reflects all that is predictable?

In addition, major factors are external and trying to predict further price only based on past price is bound to disappoint.

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#25

Post by fredliu »

Total noob question in this space (algotrading with DL algorithms).

What DL is good at, is to automate tasks that are easy for humans (e.g. telling cat from dog, understand a sentence from sound wave, translation, etc) but hard for machines without DL.

Now, if a task is even hard/borderline impossible for humans to achieve (e.g. predicting stock market with higher accuracy and consistency), why would we believe DL could do a better job than humans?

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#26

Post by DevX101 »

I'm not a purist believer in the efficient market hypothesis. BUT, I doubt there's much alpha to be gained simply from looking at price data which is widely and publicly available. Also keep in mind that markets are dynamic feedback loops so even if this model had an edge, the act of publishing this article would work against you to neutralize that edge in the future.

There's a good reason the most successful 'quantitative' trading fund, Rennaisance Technologies is so secretive and subjects its employees to a lifetime NDA/non-compete.

If there is edge in the market (and I think there are anomalies that can be traded on), I would assume it to come from correlating proprietary datasets with price. Or datasets that have a high barrier to access/analyze. One example that I'd bet still works in some industries would be counting trucks from a supplier to estimate product demand before the companies announces earnings ;)

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#27

Post by Cau5tik »

dointheatl wrote: post #24

11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.

Now, where did I put that drill?

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#28

Post by zugweg »

Do not minmax scale the data. New data outside ranges seen before will be chopped off. Do make relative (high/low), so patterns you find generalize to other domains. Don't fit complex LSTM models on tiny datasets. Somehow this is the shiny thing all starters jump on, while the most simple models are the most proven and robust. Don't just share such a model without giving a disclaimer ("I never tried this out on real data, and would not invest my own money in this solution"). This is real money, real economies, and you could horribly crash these. That is also why you don't write with authority about how to build a bridge (while you never crossed a bridge you've build before). The only reason for the "clearly showing how powerful LSTM models are" is it responding to test set samples it got wrong, and correcting with a lag. How about some evaluation, preferably on another stock than trained on?

Re: Using a Keras Long Short-Term Memory Model to Predict Stock Prices

#29

Post by SEJeff »

DevX101 wrote: post #26

I'm not a purist believer in the efficient market hypothesis. BUT, I doubt there's much alpha to be gained simply from looking at price data which is widely and publicly available. Also keep in mind that markets are dynamic feedback loops so even if this model had an edge, the act of publishing this article would work against you to neutralize that edge in the future. There's a good reason the most successful 'quanti…

Rennaisance is one large quantitative trading firm, there are other bigger and more successful ones however. I happen to work for another very large and successful quant trading firm.
Post reply on HN