This is my take on Time Series Data. Taking Sunspot data, I have tried to explain every step involved and links are provided for further study.I have analysed two more Data Set so that few more steps/processes can be shown used while dealing Time Series data.
- Time series Exploratory Analysis
- Concepts of different Averages used for approximating TS
- lag plots
- Stationarity
- Autocorrelation/Partial autocorrelation
- How to select (p,q) for ARIMA models
- AutoArima
- Forecasting a few time steps in the future.
- Many other helping operations in pandas
- One Non-stationary data is also analyzed where stationarity is achieved through the decomposition and trend is subtracted.ENZ_DATA_CENTER
- Converting predicted value back to its original scale.
- In this example Stationary is achieved through log and differenciation.
- After prediction for conveting back to its original scale cumsum and antilog steps are taken.