Time series and forecast analysis on weather data of different frequencies.
Below are some highlights of the techniques used:
Data Exploration
- Moving Average
Modelling
- Time series decomposition
- Linear regression for time series data
Data Cleaning
- Linear imputation (non-periodic time series data)
- Training and test datasets split
Modelling
- Time series decomposition
- Fourier transformation
- Linear regression for time series data
- Dynamic regression with different ARIMA error
- Lag predictors