A complete comparison between two widely used machine learning models to forecast the price of commodities shown in this article.
D. S. A. Aashiqur Reza and Tanmoy Debnath Mathematics Discipline, Khula University
A reliable forecasting model to predict future scenario of the price of two most commonly used commodity in our day to day life: wheat (retail) and rice (coarse). In this study, Seasonal Auto-Regressive Moving Average (ARIMA) and Neural Network Auto-Regressive (NNAR) were used were applied on the time series characteristic of rise and fall of prices of the commodities. Ljung-Box tests confirmed that both of the models had very good fitting and forecasting performances and NNAR model shows better fitting than ARIMA as the p-value of NNAR model were well above ARIMA model. RMSE, MAPE, MAE and MASE values were low enough to suggest with both of the models are capable of giving good predictions but comparatively NNAR model shown lower values than ARIMA model. Again NNAR model showed a better predictive performance than ARIMA model. The value of R-squared were 0.98 in NNAR model whereas it was 0.95 in ARIMA model in both cases. A pretty large training set was taken to calculate this value manually. Therefore, based on all these tests, NNAR model holds good for forecasting task of aforementioned series.
In Bangladesh, wheat and rice are two of the mostly used commodities. In recent years, the behaviors of change of price of these commodities are very mysterious and hard to predict what is going to happen in next few years as the change of weather and global pandemic strikes may cause a huge leverage in this price changes. In this study, the future behavior of price changes is forecasted based on the previous behaviors. Some useful forecasting methods used to forecast the future state by doing time series analysis. These methods analyze the patterns from the available observations and make a good prediction of future outcomes.
Applying the ARIMA model to forecast time series dataThe notion of stationarity of a series is important for applying statistical forecasting models since.
This tutorial was supposed to be published last week. Except I couldn’t get a working (and decent) model ready in time to write an article about it.
While LSTMs have become increasingly popular for time series analysis, they do have limitations. Long-short term memory networks (LSTMs) are now frequently used for time series analysis.
Time-Series Forecasting: Predicting Stock Prices Using An ARIMA Model. In this post I show you how to predict the TESLA stock price using a forecasting ARIMA model
Forecasting Time Series Data - Stock Price Analysis: Focused on forecasting the Time-series data using different smoothing methods and ARIMA in Python.