Search
× Search
Sunday, December 22, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Data with different Frequency used in classification or prediction algos?

photo

 Thomas Schlebusch, CIO at NMRQL Research Pty(Ltd)

 Monday, May 4, 2015

HI Im interested if anyone else has dealt with prediction or classification algos where the input data has varying frequency. An example of this would be daily price data vs quarterly economic or fundamental data in financial time series. I would be interested to know how others deal with this? Thanks


Print

11 comments on article "Data with different Frequency used in classification or prediction algos?"

photo

 Gregory Strzelichowski, Trader

 Monday, May 4, 2015



hi thomas, i have (in my own specific way) dealt with this issue; my approach may / may not be helpful but i'm happy to share; can you discuss off line?


photo

 Jenny Considine, Partner at Ossian Investments LP

 Tuesday, May 5, 2015



Yes, as an analyst in the natural gas, electricity, and oil industry, definitely... Also worked with different frequencies in the signal processing and medical imaging fields. Would be happy to discuss off line


photo

 Vidur (Sonny) Nanda, Founder, R&D Director, M3-IP Ltd.

 Wednesday, May 6, 2015



We have developed a real time system array for the Euro/Usd dealing with 4 time frames and 8 quantitative variable dealing with a 32 variables . Can discuss off-line


photo

 private private,

 Thursday, May 7, 2015



Yes, we can discuss this next week on a private channel. Just send me a message indicating a good time to chat next week.


photo

 Andrey Gorshkov, Algorithmic Trader, C++ Developer

 Sunday, May 10, 2015



Hi. The answer is simple: either sampling or time-weighted calculations. In the first case you convert all the data to a single time frequency. In the second case you use pairs (value, tm-delta) for all calculations and treat tm-delta as weight.


photo

 private private,

 Sunday, May 10, 2015



I do graphs with 3 moving averages, mostly 8, 13 and 21 day averages. This is done in Excel with macros. The macros do allow for other averages. Another macro looks for when these averages "converge", by coming extremely close together. The lines in a graph appear to be one on top of the other and can indicate a change in upward or downward general direction.

I am more than willing to talk of this in the open forum, unless you have a reason not to?

Thanks, David


photo

 Thomas Schlebusch, CIO at NMRQL Research Pty(Ltd)

 Monday, May 11, 2015



Thanks everyone for your help. Managed to find more than one approach. Mych appreciated.


photo

 Mark Brown mark@markbrown.com, Global Quantitative Financial Research, International Institutional Trading, Algorithmic Modeling.

 Wednesday, May 13, 2015



timeless data eliminates the need for multiple timeframes. anytime "time" is used to constrain the natural flow of data, deterioration of analysis will result. only analysis based on non-time based data will always be perfectly the same.


photo

 private private,

 Wednesday, May 13, 2015



Hi Mark,

Can you give us some examples of timeless data?

Thanks, David


photo

 private private,

 Tuesday, July 7, 2015



Repeat the values of lower frequency to the highest frequency data. Here's a visual.

Daily 1-30 1-30 1-30 1-30 1-30 1-30 1-30 1-30 …

Quarterly 1 2 3 4 1 2 3 4 …

Yearly 1 1 1 1 2 2 2 2 …

Etc.


photo

 private private,

 Tuesday, July 7, 2015



LinkedIn removed the tabs above so the formatting is off now.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association