I am making predictions on cryptocurrency / stock price data. I am using the candlestick data (High, Low, Open, Close) to estimate the close price of the next day. I am looking for the right way to normalize this kind of data. I am experienced with time series analysis / prediction but new to unbounded evolution (the price often grows exponentially). Is there a standard way to normalize this kind of data? What are the options and what are the pro's and con's?
PS. At the moment I divide every timestep with the simple moving average (SME).