Forecasting Bitcoin with technical analysis: A not-so-random forest?

This paper uses data sampled at hourly and daily frequencies to predict Bitcoin returns. We consider various advanced non-linear models based on a multitude of popular technical indicators that represent market trend, momentum, volume, and sentiment. We run a robust empirical exercise to observe the impact of forecast horizon, model type, time period, and the choice of inputs (predictors) on the forecast performance of the competing models. We find that Bitcoin prices are weakly efficient at the hourly frequency. In contrast, technical analysis combined with non-linear forecasting models becomes statistically significantly dominant relative to the random walk model on a daily horizon. Our comparative analysis identifies the random forest model as the most accurate at predicting Bitcoin. The estimated measures of the relative importance of predictors reveal that the nature of investing in the Bitcoin market evolved from trend-following to excessive momentum and sentiment in the most recent time period. (c) 2021 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

Automated summary

This paper uses data sampled at hourly and daily frequencies to predict Bitcoin returns. Various advanced non-linear models based on popular technical indicators are considered to observe the impact of forecast horizon, model type, time period, and choice of inputs on the forecast performance. Findings show weak efficiency of Bitcoin prices at hourly frequency, while technical analysis combined with non-linear forecasting models dominate on a daily horizon. The random forest model is identified as the most accurate predictor for Bitcoin. The study also reveals the evolution of investing in the Bitcoin market from trend-following to excessive momentum and sentiment in recent time period.