The forecasting of agriculture commodity price is mainly divided into two parts: structural and nonstructural. The structural methods [5] specifically recollect the deliver demand ratio. Computationally, it's miles very tough to estimate the customers’ desires and the manufacturing of that specific crop for developing nations. The nonstructural methods [6] may be labeled as statistical method [7, 8] and machine gaining knowledge of method. For nonstructural methods, historical records are accumulated as time collection data. The time series records can be of linear or nonlinear in nature. There are various methods [9] for forecasting primarily based on time collection information. short term forecasting The ANN is a better opportunity than the statistical model for nonlinear time collection data [10, 11].

Some studies research were carried out in forecasting of agriculture commodities charge in developing international locations such as India. According to [12, 13], there are a few special features of the ANN which include nonlinearity, adaptability, and mapping techniques providing robust assist for the usage of the ANN as a very good forecasting version.

In [14], the ARIMA version and time postpone neural network (TDNN) are for time series forecasting of agriculture commodity rate. They concluded that the neural network model completed higher due to nonlinear nature of time series information. Finally, they presented a hybrid model for forecasting. Surprisingly, the hybrid version turned into less green than the ANN for soybean statistics and extra efficient for mustard.

According to the paintings suggested in [15], the neural network is offered which is a very good opportunity for “brief time period” forecasting, at the same time as the Box–Jenkins approach performs better for terribly brief-time period forecasting. They also mentioned that the neural community with out a hidden layer can paintings just like the Box–Jenkins technique.

Work offered in [16] used the assist vector device for forecasting of economic time series information to perform higher in terms of efficiency in contrast with the back propagation neural network.

In [17], the authors supplied the ANN approach for multivariate time collection facts. accounts receivable management They used the dataset of flour price of 3 towns, and based on training and checking out consequences, they concluded that the ANN model can well be used for forecasting.

In [18] too, the ANN version is used for electric load forecasting. They used the traits of the ANN to research from the connection most of the beyond records, modern, and future temperature. Based on the checking out facts, the end result become very exceptional.