As a new type of digital currency,
Bitcoin
is considered as “future gold” by various scholars. Therefore, this study considers
Bitcoin
and gold as a group of hedging assets to conduct investment research and it also discusses the investment rules between
Bitcoin
and gold: prediction of the rise and fall of
Bitcoin
, comparison of the characteristics of
Bitcoin
and gold, and the impact of the transaction procedures of
Bitcoin
and gold on the final trading results, and formulates trading strategies through optimization algorithms. Then, four machine learning algorithms, i.e., LSTM, BP neural network, Adaboost, and Bagging, are introduced to predict the rise and fall of gold and
Bitcoin
the next day, and then, the entropy weight method is used to synthesize four predicted results to ensure the robustness of the predicted results. To establish the optimal trading strategy, this study considers the maximum expected return as the goal to develop a single-objective optimization model and historical five-day price volatility as a risk factor. In this study, ant colony, simulated annealing, and genetic algorithms are used to solve the single-objective optimization model. Finally, we conclude that
Bitcoin
, similar to other financial assets, e.g., gold, is sensitive to shocks and volatile and possesses a relatively quiet cycle. When
Bitcoin
has an asymmetric impact,
Bitcoin
and gold can equally treat transactions.
抄録全体を表示