Project 2: Energy production prediction using Time series Models
-
The project aimed to predict energy production using time series models, which are statistical models used to analyze data over time to make forecasts or predictions.
-
Collected and analyzed energy production data over time, which was then cleaned and preprocessed to prepare it for modeling.
-
Preprocessed the data by removing trends i.e., subtracting the rolling mean and decomposition. Autocorrelation analysis was also done to detect the patterns and check for randomness.
-
Implemented time series modeling techniques i.e., autoregressive (AR) model, moving average model, autoregressive integrated moving average (ARIMA), and Facebook Prophet Model (basic and tuned).
-
Feature engineered to extract relevant features such as seasonality and trend in the data to improve the accuracy of the models.
