BITCOIN OPENING VALUES ANALISYS
Palavras-chave:
Bitcoin, crypto-currency, Deep Learning, Long short-term memory networksResumo
A crypto-currency is a decentralized exchange means that uses blockchain technology, which
is a type of distributed log book operated on a point-to-point network of thousands of computers, where
everyone has an equal copy of the entire transaction history, preventing a central entity from promoting
changes to the registry or to the software unilaterally, and from cryptography to ensure the validity of
transactions and the creation of new currency units. Bitcoin, the first decentralized digital currency in
history, was created in 2009. The idea with Bitcoin was to replicate the properties of physical money in
a digital environment, the digitization of money in the fullest sense of the term. Since then, many other
crypto-coins have been created, and the challenge has been raised to discover which crypto-coins are
most promising to invest. Investing in the financial market involves buying and selling financial assets,
and every investor expects to find standards in the market that allow him to buy low value’s assets and sell
high value’s assets. The objective of this work is to propose a deep learning method to learn the behavior
of increasing or decreasing the opening values of each day of the Bitcoin crypto-currency in an interval of
up to 20 days. We used Bitcoin purchase values from April 2013 to February 2018. Different forecasting
strategies were studied. The experimental results showed relevance in the price trend of Bitcoin.