A Model for Quantifying Investment Decision-making Using Deep Reinforcement Learning (PPO Algorithm)

Abstract

Bitcoin has nowadays shown its importance in finance systems. With great interest to all kinds of investors in the financial sector, it is important to analyse the relationship between Bitcoin and gold. We consider bitcoin and gold as two stocks and calculate the correlation between bitcoin and gold. By introducing the calculation of dynamic penalty coefficient, the problem of dual-stock portfolio investment is transformed into the problem of single-stock purchase investment, which greatly reduces the difficulty of feature engineering and model application. In terms of decision-making model, deep reinforcement learning (PPO algorithm) is used to make quantitative investment decisions. Therefore, we use the expected data in SLTM as the input data of deep reinforcement learning, and combine it with deep reinforcement learning for training. Compared with machine learning to quantify investment decisions, after a period of training, the accuracy rate has improved by 10.038%.