Many professional traders earn their profits not only by utilizing current market state, but also by understanding when not to trade. Excess trading leads not only to losing trades but also to increased commissions which may massively impact your profits (especially in intraday, HFT space). Meta-Labelling technique is a financial machine learning technique designed to filter-out false positives, reduce transaction costs and overtrading. On this lecture we will discuss:
1. The architecture and rationale behind Meta-Labelling model.
2. How Meta-Labelling solve the issue of precision/recall optimisation in algorithmic trading.
3. Will show how Meta-Labelling model improves the strategy performance on the example of AR(3) process.
4. Discuss how to size your postions with Kelly criterion and using probabilities form Meta-Labelling model.
When? September 6, 19:00 (Kyiv time)
100% of the funds from ticket sales will be transferred to the charity fund Твоя Опора for the purchase of devices for children’s resuscitation, which save the lives of children who suffered as a result of russian aggression.