Portfolio investment decision model based on price forecasting model and risk assessment model

Abstract

Investors often use gold and Bitcoin to reduce the risk of financial investments by combining them due to their low correlation and different dynamic correlation. Therefore, it is essential to study how a combination of the two can be invested to maximize returns. In this paper, we combine the BP neural network model and the gray correlation analysis model from the future unpredictability and risk quantification of portfolio investment. We want to use the model to construct a price prediction model and risk assessment model to build a model for portfolio investment decisions based on past-day price flow only. To be specific, this paper first builds a price prediction model based on a BP neural network prediction model. At the same time, the prediction needs to be performed in a cycle, and the training set of the neural training model is increased each time. Each prediction is recorded for the next day's data and aggregated to obtain the daily price flow prediction. Next, this paper builds a risk assessment model through gray correlation analysis. In order to measure the riskiness of the daily investment, this paper selects three indicators, namely, forecast price accuracy, recent price trend, and price volatility. It uses gray correlation analysis to score them quantitatively and then gives the daily investment risk assessment results. Finally, a sensitivity analysis was performed to determine the model's sensitivity to transaction costs by varying the transaction costs of gold and bitcoin. It was found that as the transaction costs become more extensive, the number of transactions becomes smaller and the total assets acquired become smaller.