STOCK TRADING OF JII ISLAMIC STOCKS USING DEEP REINFORCEMENT LEARNING
From the perspective of devout Muslims, they are restricted to investing only in assets that are not contrary to Islamic principles. However, it is challenging to gain optimal return with fewer alternatives in Islamic stocks. To address this challenge, the potential of Deep Reinforcement Learning...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Subjects: | |
| Online Access: | https://digilib.itb.ac.id/gdl/view/57416 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | From the perspective of devout Muslims, they are restricted to investing only in assets
that are not contrary to Islamic principles. However, it is challenging to gain optimal return
with fewer alternatives in Islamic stocks. To address this challenge, the potential of Deep
Reinforcement Learning (DRL) is explored to optimize the stock trading. Stock trading task is
modeled as a Markov Decision Process (MDP) problem, due to its stochastic and interactive
nature. Then the trading objective is a problem of maximization. The DRL agents used are
actor-critic algorithms, namely A2C, DDPG, and PPO. The selected portfolio consists of 30
most liquid Islamic stocks in Indonesia that constitutes JII index. Finally, the performance is
compared between the algorithms, and against LQ45 index as the benchmark that consists of
the 45 most liquid conventional stocks in the Indonesian Stock Exchange (IDX). The result
shows that trading on Islamic stocks from Jan 2019 to Dec 2020 using the DRL agents can
outperform the benchmark index of conventional stocks. |
|---|