Purpose

Since the issuance in 2009, the digital currency has enjoyed an increasing popularity and has become one of the most important options for global investors. The purpose of this paper is to propose a hybrid model ( KDJ–Markov chain) which integrates the advantages of the stochastic index (KDJ) and grey Markov chain methods and provide a useful decision support tool for investors participating in the digital currency market.

Design/methodology/approach

Taking Litecoin's closing price prediction as an example, the closing prices from May 2 to June 20, 2017, are used as the training set, while those from June 21 to August 9, 2017, are used as the test set. In addition, an adaptive KDJ–Markov chain is proposed to enhance the adaptability for dynamic transaction information. And the paper verifies the effectiveness of the KDJ–Markov chain method and adaptive KDJ–Markov chain method.

Findings

The results show that the proposed methods can provide a reliable foundation for market analysis and investment decisions. Under the circumstances the accuracy of the training set and the accuracy of the test set are 76% and 78%, respectively.

Practical implications

This study not only solves the problems that KDJ method cannot accurately predict the next day's state and the grey Markov chain method cannot divide the states very well, but it also provides two useful decision support tools for investors to make more scientific and reasonable decisions for digital currency where there are no existing methods to analyze the fluctuation.

Originality/value

A new approach to analyze the fluctuation of digital currency, in which there are no existing methods, is proposed based on the stochastic index (KDJ) and grey Markov chain methods. And both of these two models have high accuracy.

中文翻译：

---

预测数字货币的价格趋势：融合随机指标和灰色马尔可夫链方法的混合模型

目的

自2009年发行以来，数字货币已越来越受欢迎，并已成为全球投资者最重要的选择之一。本文的目的是提出一种混合模型（KDJ-马尔可夫链），该模型融合了随机指标（KDJ）和灰色马尔可夫链方法的优势，并为参与数字货币市场的投资者提供了有用的决策支持工具。

设计/方法/方法

以莱特币的收盘价预测为例，将2017年5月2日至6月20日的收盘价作为训练集，而将2017年6月21日至8月9日的收盘价作为测试集。此外，提出了自适应KDJ-马尔可夫链，以增强动态交易信息的适应性。并验证了KDJ-Markov链方法和自适应KDJ-Markov链方法的有效性。

发现

结果表明，所提出的方法可以为市场分析和投资决策提供可靠的基础。在这种情况下，训练集的准确性和测试集的准确性分别为76％和78％。

实际影响

这项研究不仅解决了KDJ方法无法准确预测第二天的状态以及灰色马尔可夫链方法无法很好地划分状态的问题，而且还为投资者提供了两个有用的决策支持工具，使投资者可以做出更科学合理的数字决策。没有可用的分析波动的方法的货币。

创意/价值

基于随机指数（KDJ）和灰色马尔可夫链方法，提出了一种新的分析数字货币波动的方法，该方法没有现有的方法。而且这两个模型都具有很高的精度。