From the course: Python for Time Series Forecasting

Unlock this course with a free trial

Join today to access over 24,800 courses taught by industry experts.

Summary: Important steps to consider in ARIMA modeling

Summary: Important steps to consider in ARIMA modeling - Python Tutorial

From the course: Python for Time Series Forecasting

Start my 1-month free trial Buy for my team

Summary: Important steps to consider in ARIMA modeling

“

- [Instructor] In statistical forecasting with ARIMA, where we develop a mathematical model to forecast the future values based on the historical regular intervals of a time series, because they're following a fixed number of intervals, in this case, the months. We see that there is a pattern that we must model with the ARIMA class coming from the statsmodels library. The most important thing is to configure these parameters, p, d, and q, where the first step is to differentiate so that we can certify the constant mean and variance over time. And this was corroborated by the Dickey-Fuller test, which in this case, we are looking for p-values which are lower than 0.05. It's the only one in this diagnostics of modeling time series. Then we went to the partial autocorrelation function and also the autocorrelation function, where we get these lollipops, which are the lags that say how correlated the current information with one step prior or two step priors. We get these values to be…

Contents