I'm new to machine learning and have made some experience with some basic dataset and now i'm exploring BTC data. I'm asking is correct to shuffle data when the data are BTC, ETH or any other crypto? What i mean is that the price of a stock or crypto increase or decrease in every candle with a specific time-frame. Does have sense to shuffle by removing the sequentiality of the candles?
Asked
Active
Viewed 39 times
2 Answers
0
I don't think it's relevant to remove the time aspect for the purpose of predicting a price. Indeed, if you want to predict a price, you need to look at previous prices and therefore have a temporal aspect to be able to detect trends. Moreover, if you make a program based on ML, it is important to keep the notion of time in order to be able to make "windows" to determine the price at d+1 or other. If anyone disagrees, or would like to add something, it would be a pleasure So don't shuffle
0
Shuffling data will only be good for crypto and BTC if there is no relation between the sequence of events. (That is, the order in which the price changes occur) which will negatively affect the future price, making it harder for the model to learn meaningful patterns and accurately predict future prices.
I_Al-thamary
- 101