<![CDATA[Forecasting is a time series data analysis technique for predicting future data by obtaining patterns of change in past data. Exponential smoothing, AutoRegressive Integrated Moving Average (ARIMA), and Box-Jenkins are common forecasting algorithms for linear time series data. Meanwhile, models such as Artificial Neural Networks (ANN), Fuzzy, and others are frequently utilized for nonlinear time series data. One of the most generally used model selection procedures is to evaluate each model that has been trained in time series data learning and then used to predict the model's performance, and then allow the forecaster determine if the model is acceptable or choose the best model from a list of candidates. Forecasts created with the best model, on the other hand, rarely produce generalized outcomes for the full data set. As a result, it's crucial to put the results of the learning training to the test. The ensemble method is employed instead of learning from a large number of models. The objective of this research is to apply ANN and the Ensemble Approach to optimize a forecasting model. When forecasting with a neural network, the ensemble approach is used to limit the occurrence of over fitting so that the resulting model can beat individual NN models and be consistent in lowering mistakes.]]>
Copyrights © 2022
