Analysis of Trading Strategy based on ARIMA and Dynamic Programming Model

10.6919/ICJE.202205_8(5).0060

Zhiduo Wang；Zixing Liu；Shuai Yin

Lagrange Interpolation ； ARIMA Model ； Dynamic Programming Model

International Core Journal of Engineering

8卷5期（2022 / 05 / 01）

471 - 480

英文

Market traders obtain the maximum value by buying and selling two volatile assets, gold and bitcoin. This paper analyzes the historical data of each day to establish different asset value prediction models and investment transaction planning models, so as to give the market The best investment strategy for traders. First, use Lagrange interpolation method to supplement the missing data in gold, predict the future value by using the effective real data of gold and bitcoin, and select the previous 30-day data of gold and bitcoin as the training set respectively, and establish an ARIMA-based gold and bitcoin model. The Bitcoin prediction model predicts the value of the 31st day, and then predicts the value of different assets on the 32nd day through the real data of the first 31 days, so as to recursively give the predicted value of gold and Bitcoin in the past five years. Secondly, introduce the Sharpe ratio that has both risk and profitability, and use the predicted Sharpe ratio of the n+1 day as the objective function, the daily profit value is greater than or equal to the transaction commission, and invest in the cash, gold and bitcoin pairs after the transaction. The proportion of total assets is a non-negative number, these four are set as constraints, and the change in the proportion of gold and bitcoin is set as a decision variable, and an asset investment transaction model based on dynamic programming is established to make reasonable decisions on daily transactions, and get the results for each day. Cash, gold, bitcoin holding rates to get the highest benefits.

1. BERRUT JP, TREFETHEN LN. Barycentric Lagrange interpolation[J]. SIAM Review,2004,46(3):501-517.
   連結：
2. CHEHELGERDI-SAMANI, MARYAM, SAFI-ESFAHANI, FARAMARZ. PCVM.ARIMA: predictive consolidation of virtual machines applying ARIMA method[J]. Journal of supercomputing,2021, 77(3): 2172-2206. DOI:10.1007/s11227-020-03354-3.
   連結：
3. SRIVASTAVA AKHILESH KUMAR, SRIVASTAVA ANAND, SINGH SIDDHARTHA, et al. Design of Machine-Learning Classifier for Stock Market Prediction[J]. SN Computer Science,2021,3(1). DOI: 10.1007/s42979-021-00970-5.
   連結：
4. DELSOLE, TIMOTHY, TIPPETT, MICHAEL K.. Correcting the corrected AIC[J]. Statistics & Probability Letters,2021,173. DOI:10.1016/j.spl.2021.109064.
   連結：
5. An analysis of the Hypervolume Sharpe-Ratio Indicator[J]. European Journal of Operational Research, 2020,283(2):614-629. DOI:10.1016/j.ejor.2019.11.023.
   連結：