Showing 1–1 of 1 results for author: Bekezin, N

DeepFolio: Convolutional Neural Networks for Portfolios with Limit Order Book Data

Authors: Aiusha Sangadiev, Rodrigo Rivera-Castro, Kirill Stepanov, Andrey Poddubny, Kirill Bubenchikov, Nikita Bekezin, Polina Pilyugina, Evgeny Burnaev

Abstract: This work proposes DeepFolio, a new model for deep portfolio management based on data from limit order books (LOB). DeepFolio solves problems found in the state-of-the-art for LOB data to predict price movements. Our evaluation consists of two scenarios using a large dataset of millions of time series. The improvements deliver superior results both in cases of abundant as well as scarce data. The… ▽ More This work proposes DeepFolio, a new model for deep portfolio management based on data from limit order books (LOB). DeepFolio solves problems found in the state-of-the-art for LOB data to predict price movements. Our evaluation consists of two scenarios using a large dataset of millions of time series. The improvements deliver superior results both in cases of abundant as well as scarce data. The experiments show that DeepFolio outperforms the state-of-the-art on the benchmark FI-2010 LOB. Further, we use DeepFolio for optimal portfolio allocation of crypto-assets with rebalancing. For this purpose, we use two loss-functions - Sharpe ratio loss and minimum volatility risk. We show that DeepFolio outperforms widely used portfolio allocation techniques in the literature. △ Less

Submitted 27 August, 2020; originally announced August 2020.