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

111–120 of 178 posts

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

#111

Post by JumpCrisscross »

ttobbaybbob wrote: post #91
throwawaymath wrote: post #81

Earlier quoted context omitted.

I'm pretty sure that lifetime NDA is a myth. It gets repeated very often (like a lot of the Renaissance mythos) but I haven't seen evidence for it, and on its face it sounds flagrantly unenforceable. It could be that the compensation alone substantially reduces turnover. Likewise once they've hit "their number", employees may decide to go into tech or philanthropy instead of a competing firm. For what it's worth, Ren…

The NDA/non-compete may not enforceable, but if Renaissance believes employees with valuable IP are being poached, they will sue and make it very expensive to hire, then defend, ex-rentec employees. https://www.marketwatch.com/story/renaissance-millennium-set...

They’re totally up front about this requirement, too, when making offers. (Source: I turned one down.) You are to move to Long Island, buy a house, buy into the employees-only fund and work there for the rest of your financial career.

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

#112

Post by martinlaz »

alehul wrote: post #73
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…

Rentech and the other more secretive top quant firms such as TGS and PDT use nearly exclusively public data. Funds like Two Sigma that haven't had as good (or scalable) returns are actually the ones that focus on novel data sources. In reality, everyone in this industry is drowning in data, and the real edge comes from learning how to more efficiently parse and analyze data rather than acquiring more of it. The top p…

These are bold claims with nothing to support them. Given the claimed secrecy, I don't see how you could possibly know all that, unless you are (or have been) working for Rentech, or one of the others. Is that the case?

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

#113

Post by vl »

goodside wrote: post #93

This is pure nonsense. This isn’t even the right way to begin thinking about this as a forecasting task — the target series should be log-normal returns, not raw asset price. The performance of this model is laughably bad, which is probably why he spends zero time evaluating its effectiveness. You could trivially get better forecasts than this by naively repeating the last-observed price . This isn’t ML. It’s cargo-c…

Even more so since test data can't be from the same time range: i.e. for time series you need to split train/test by date, not randomly, otherwise your model just memorizes the series.

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

#114

Post by jpeterson »

jwiley wrote: post #102
goodside wrote: post #93

This is pure nonsense. This isn’t even the right way to begin thinking about this as a forecasting task — the target series should be log-normal returns, not raw asset price. The performance of this model is laughably bad, which is probably why he spends zero time evaluating its effectiveness. You could trivially get better forecasts than this by naively repeating the last-observed price . This isn’t ML. It’s cargo-c…

Can you clarify why "the target series should be log-normal returns" is important or provide a pointer for more information?

Investment markets operate on relative gain, not absolute gain. E.g.: if you invest in a stock and it gains $5, this would be a great return for a $1 stock but a poor one for a $1000 stock, so the absolute gain doesn't mean anything on its own. A 5% return always means that you've gained 5% on your investment.

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

#115

Post by WrtCdEvrydy »

JumpCrisscross wrote: post #111
ttobbaybbob wrote: post #91

Earlier quoted context omitted.

The NDA/non-compete may not enforceable, but if Renaissance believes employees with valuable IP are being poached, they will sue and make it very expensive to hire, then defend, ex-rentec employees. https://www.marketwatch.com/story/renaissance-millennium-set...

They’re totally up front about this requirement, too, when making offers. (Source: I turned one down.) You are to move to Long Island, buy a house, buy into the employees-only fund and work there for the rest of your financial career.

LOL, No.

How do you even enforce that? If I get a job somewhere else, there's no reason for me to disclose where I'm going.

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

#116

Post by colinchartier »

jwiley wrote: post #102
goodside wrote: post #93

This is pure nonsense. This isn’t even the right way to begin thinking about this as a forecasting task — the target series should be log-normal returns, not raw asset price. The performance of this model is laughably bad, which is probably why he spends zero time evaluating its effectiveness. You could trivially get better forecasts than this by naively repeating the last-observed price . This isn’t ML. It’s cargo-c…

Can you clarify why "the target series should be log-normal returns" is important or provide a pointer for more information?

I think their point is that predicting that "the stock will go up by 1$" or "the stock will go down by 1$" is worse than "the stock will go up by 0.05%" and "the stock will go down by 0.05%" because of this little paradox:

50$ increase from 50$ is 100% increase 50$ decrease from 100$ is 50% decrease

e.g., if the model finds 50$ increase/decreases, that actually corresponds to very different wealth changes

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

#117

Post by argonaut »

jermaustin1 wrote: post #64
DevX101 wrote: post #53

Earlier quoted context omitted.

My understanding is that legally, the contract says something reasonable, like a year. But the firm will use it's considerable resources to impede your career if you attempt to work in mathematical trading after that year's up. On the flip side, employees are compensated very well, and the turnover is very low.

I'd assume turn over is low if by leaving they now have to change entire career paths. I'd put up with a lot of shit to not have to learn a brand new trade.

They don't. They're all qualified for machine learning or data scientist roles in any area of tech. That's a huge field.

Most of them are also qualified for general software engineering roles; some of the people with more theoretical backgrounds might not be.

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

#118

Post by aggronn »

derefr wrote: post #105
aggronn wrote: post #74

Earlier quoted context omitted.

The problem I see with that hypothesis is that the asset pricing must eventually be tied to actual performance--while it may be self-reinforcing to some extend, if it _is_ fundamentally wrong, there will be a reckoning and an adjustment after a high value stock goes bankrupt, for example.

But the stock market isn’t just a bunch of people placing speculative bets about company performance; it’s a bunch of people placing speculative bets about company performance by buying shares of the companies . If the market is irrational, it can actually prop up companies that would otherwise go bankrupt.

I suppose they company could continuously issue stock in this case.

But in this world, it actually makes sense for those companies to stop doing their normal business and just go into the business of selling their shares.

This reality sounds absurd, but you could argue that BTC market is there. Enough of the market thinks that "always buy" is a good investment, regardless of the real-life value of the asset. Or maybe, less controversially, gold is that market. Any asset that is always increasing in value and not related to the real value of that asset is just a store of value.

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

#119

Post by mmq »

samfisher83 wrote: post #58

All these LSTM Stock Market tutorials seem to be a variation of this guys tutorial which he did a while ago: https://github.com/jaungiers/LSTM-Neural-Network-for-Time-Se... Except Jakob explains the folly of the method. Whereas the author of this piece says the following: From the plot we can see that the real stock price went up while our model also predicted that the price of the stock will go up. This clearly show…

I made this code https://github.com/mouradmourafiq/tensorflow-lstm-regression at least 6 months earlier than the first push in the github repo you mentioned, and I am sure many people did some version before that. A lot of people sometimes just need a reason to play with some technology. I assume that OP used stock market to learn about LSTMs.

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

#120

Post by SyneRyder »

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.

12:50 Press Return

(For anyone who didn't get the reference, you have an excellent movie & coding soundtrack to catch up on: https://www.youtube.com/watch?v=ShdmErv5jvs)

Post reply on HN