I have started the TimeSeries project about predicting Bitcoin BTC price.
First I am going to try a traditional method: ARIMA. Will it work? 🤔
If you don't know what it is 🧵👇
datascience artificialintelligence machinelearning
It is composed of 3 elements: 1️⃣
AR: Auto-Regressive 2️⃣
I: Integrated 3️⃣
MA: Moving Average
Let's introduce them... 👇
[ AR ]-IMA
🔎It takes into account "p" previous values, in my case previous BTC prices.
🔮To make the prediction it relies on the previous prices.
AR-[ I ]-MA
🔎It differentiates the values "d" times.
🔮The purpose of this is to make the time series stationary, which is a requirement for making predictions with this method.
ARI-[ MA ]
🔎It takes into consideration "q" previous residuals (actual value - prediction).
🔮To make the prediction it relies on how off the previous predictions were.
🤔But how to choose those three parameters "p", "d" and "q" ?
Soon more on this 😉
If you don't know what project I am talking about 👇👇👇👇This post is based on this twitter thread.