**5. Experimental Results and Discussion**

This section sets out our three research questions. We will first examine the impact of noise and trend effects on the correlation between cryptocurrencies as well as their corresponding topological structure. Then, we observe the evolution of the structure according to the levels of granularity. Finally, the results from these two experiments will be used to construct the right network structure. Consequently, the corresponding community structure is identified, which is used to learn the investment decisions of crypto investors during the COVID-19 pandemic.

We note that all calculations in our study are implemented using *Python* programming language (version 3.7.14, designed by Guido van Rossum, Centrum Wiskunde & Informatica (CWI), The Netherlands). Regarding network-related calculations (e.g., network construction and network-involved metrics), we utilize the *networkx* (https://networkx.org/) package incorporated into Python.
