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.

Accern Launches Machine Learning Sentiment Data Feed for Retail Quants

photo

 Kumesh Aroomoogan, Co-Founder and CEO at Accern

 Tuesday, July 28, 2015

http://www.digitaljournal.com/pr/2625931


Print

6 comments on article "Accern Launches Machine Learning Sentiment Data Feed for Retail Quants"

photo

 Kumesh Aroomoogan, Co-Founder and CEO at Accern

 Sunday, August 2, 2015



$199 is actually our pilot program. For 100 days you will have access to the feed for $199 so basically 3 months.

You can sign up for the pilot and my team will give you the API access token along with the 3 years of historical data to backtest.


photo

 tasos grivas, Implementation Consultant at Razor Risk Technologies, FRM

 Sunday, August 2, 2015



but the backtesting will take place with data that might have been used to (over)-fit your indicator... would it be possible to provide some limited type of access to your system for a month just for forward testing?


photo

 private private,

 Monday, August 3, 2015



Backtesting a strategy is useless for a simple reason that the strategy was (probably) created to fit that data in a first place.


photo

 Kumesh Aroomoogan, Co-Founder and CEO at Accern

 Monday, August 3, 2015



@Tasos: In sample data goes from August 2012 - August 2014. Out of sample data is from August 2014 to present. The pilot program gives you access to the data for 3+ months for $199. That around $66/mo which will let you get comfortable with the data feed and live test it with whatever strategies you decide to build.

@Germann: You would first create a strategy and run the backtest on the 2 years of in sample data (August 2012 - August 2014). If the strategy performs well, continue it on the out of sample data (August 2014 - now). If it still performs well, run it on our live feed. That is one of the conventional approaches to avoid any overfitting while backtesting and having confidence in the performance of your strategy.


photo

 Kaustabh Ray, Equity Research Technologist and Systems Architect

 Monday, August 3, 2015



I agree with @Germann and I definitely think back testing does not help, not with your kind of data.

The only way to properly test is forward testing as suggested by @Tasos.


photo

 Graeme Smith, Investment Manager at The Tourists Portfolio

 Tuesday, August 4, 2015



I might be interested in checking it out. But I would suggest that if you are confident in your data you should be considering a free trial for three months. A lot more people will bite at that, and if they are impressed enough to subscribe you'll make the initial trial price in the next two months. Having a paid trial suggests that you don't have much confidence and see the trial as a revenue generator.

At the least, are you able to supply the history for free to determine if it has potential?

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