Predicting crypto asset price bursts using the correlation tensor of trading networks - towards secure crypto assets and the digital economy -

Cryptoassets are becoming an integral part of the digital economy. RIKEN and Kyoto University are collaborating on theoretical research to detect anomalous events such as money laundering and fraud, and to predict price bursts, by making full use of the mathematical science of networks. A research group led by Yuichi Ikeda (Professor at Kyoto Univ.), Abhijit Chakraborty (Assistant Professor at Kyoto Univ. and Visiting Research Scientist at RIKEN iTHEMS) and Tetsuo Hatsuda (RIKEN iTHEMS), has developed a novel method to analyse the spectrum of the correlation tensor corresponding to a trading network for cryptographic assets, and has developed a new method for predicting price bursts. They found that the maximum singular value of the tensor shows a significant negative correlation with the price of cryptoasset prices. Using this finding, the group gained the prospect of providing an early indicator of price bursts.

For details, please refer to Kyoto University's press release article at the related links below.