Join the 80,000 other DTN customers who enjoy the fastest, most reliable data available. There is no better value than DTN!

(Move your cursor to this area to pause scrolling)




"And by the way, have to say this. I love the IQFeed software. It's rock solid and it has a really nice API." - Comment from Thomas via RT Chat
"The service is great, I see a noticeable improvement in my volume profiles over [broker]'s data feed" - Comment from Larry
"You have an excellent product !!!!!!" - Comment from Arely
"Everything is working great with the API. I love it." - Comment from Calvin
"With HUGE volume on AAPL and RIMM for 2 days, everyone in a trading room was whining about freezes, crashes and lag with *******, RealTick, TS and Cyber. InvestorRT with IQFeed was rock solid. I mean SOLID!" - Comment from Public IRC Chat
"I use IQ Feed, Great stuff as far as data analysis information, storage and retrieval is concerned." - Comment from Public Forum
"Thanks for the great product and support. During this week of high volume trading, my QuoteTracker + IQ Feed setup never missed a beat. Also, thanks for your swiftness in responding to data issues. I was on ******* for a few years before I made the switch over early this year, and wish I had done it a long time ago." - Comment from Ken
"Interactive Brokers tick data was inconsistent, so I have switched to using DTN exclusively. It is great to no longer have to worry about my datafeed all day long." - Comment from Philippe
"Thanks for following up with me. You guys do a great job in tech support." - Comment from Phelps
"IQFeed version 4 is a real screamer compared to anything else I have seen." - Comment from Tom
Home  Search  Register  Login  Recent Posts

Information on DTN's Industries:
DTN Oil & Gas | DTN Trading | DTN Agriculture | DTN Weather
Follow DTNMarkets on Twitter
DTN.IQ/IQFeed on Twitter
DTN News and Analysis on Twitter
»Forums Index »General Discussion »Trading Tips, Tricks and Ideas »Machine learning and tick-by-tick data
Author Topic: Machine learning and tick-by-tick data (2 messages, Page 1 of 1)

keohir808
-Interested User-
Posts: 6
Joined: Sep 16, 2019


Posted: Jan 30, 2022 08:45 PM          Msg. 1 of 2
There seems to be a lack of research regarding the use of tick-by-tick data as input to machine learning models. Has anyone experimented with machine learning and tick-by-tick data? I’ve trained an LSTM with about 2 years worth using tick-by-tick data with DTN which fit a certain criteria such as float, volume, price. The result is a model with around 69.9% accuracy. A naïve model which predicts that the bid price will be the same price as the last tick has an accuracy of around 65%. I’m wondering if I can increase my model’s accuracy through feature engineering. Could anyone share research papers regarding machine learning and tick-by-tick data? Does anyone have any insight regarding data transformations that can be applied to financial data which may result in increased accuracy if used as a feature in machine learning models?


Interests & Tools: Machine Learning, Neural Networks, Deep Learning, Python, Java, Trading, Small Caps, Interactive Brokers.
Edited by keohir808 on Jan 30, 2022 at 08:50 PM

taa_dtn
-DTN Evangelist-
Posts: 150
Joined: May 7, 2004


Posted: Jan 31, 2022 11:10 AM          Msg. 2 of 2
Yes, I experimented with this a few years ago. Your questions are relevant and insightful, but I don't have much useful information to offer in reply.

I haven't seen many published papers on the subject in recent years. Take that with a grain of salt, though, because I'm not looking actively enough. Possibly if the technique has been applied successfully, it hasn't been discussed in public for the obvious reasons. Hopefully someone else will reply with better information.

In general, I hit the same roadblocks you did. It's hard to choose the right network architecture (financial data isn't statistically stationary, so I wasn't able to design either recurrent or convolutional networks that were consistently successful). Raw tick-by-tick data has so much variability along so many dimensions that I suspect feature engineering is necessary, but that's a major research project in its own right. Techniques currently being used for natural language processing are probably where I'd start if I were to look at this again today.

Possibly the most fundamental problem I ran into is that it doesn't seem workable to use a scalar value to measure outcomes, so anything based on simple gradient descent is problematic. I think a practical outcome measurement must be at least three-dimensional -- it needs to include return, risk, and capital management. Arguably more, but the need for those three is easy to understand.
 

 

Time: Wed March 22, 2023 7:37 PM CFBB v1.2.0 47 ms.
© AderSoftware 2002-2003