| Title |
Kriptovaliutų ir internetinės informacijos sąryšių tyrimas taikant duomenų gavybos metodus / |
| Translation of Title |
Relation between cryptocurrency and online information analysis applying data mining methods. |
| Authors |
Petrošius, Giedrius |
| Full Text |
|
| Pages |
60 |
| Keywords [eng] |
cryptocurrency ; LSTM ; dynamic topic modeling |
| Abstract [eng] |
Bitcoin is a trending digital currency that is used worldwide to make online payments. Bitcoin and other cryptocurrencies have consequently become an investment vehicle in itself and are traded in a way similar to other open currencies. The aim of this thesis to use data mining techniques on online information in order to find cryptocurrency market patterns. To accomplish thesis goal dynamic topic modeling was applied to 1,37 million comments in order to find hidden topics evolutions over a period of 4,3 years from January 2014 to March 2018. The research is also concerned with predicting the price of Bitcoin using online information. The task was achieved with Long-Short Memory (LSTM) neural network which highest accuracy is 112,65 RMSE and MAPE – 0,65%. Based on LSTM model Bitcoin automated trading strategy was developed which decisions sell and bitcoins was 94% accurate. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2018 |
