CROP PRODUCTION UNDER DIFFERENT CLIMATIC CONDITIONS BY ANALYZING AGRICULTURAL DATA USING MULTIPLE LINEAR REGRESSION, WINTER HOLT, AND ARTIFICIAL INTELLIGENCE

Abstract

The article deals with the prediction of productivity dynamics of agricultural products based on exponential smoothing and optimization using Holt-Winters, multiple linear regression and ANN. Scaling the data is one of the preprocessing steps of the optimization algorithms in the dataset. As we know, most methods of ANN make decisions depending on their underlying data sets. Often, the algorithms calculate the distance between data points to draw better conclusions from the data. The effectiveness of the optimization methods is measured by the average percentage error (MAPE). According to the data calculation results, the Holt-Winter method MAPE value in prediction is 26.129 (), 297 (), 60.384 (), 93.6 (), 52.9 (), and the smallest MAPE value in multiple linear regression method is 0.28, MAPE value for ANN method is 128 Evolved. Considering the level of MAPE, the MAPE value in the ANN method decreased from 685.6 to 93.6 in comparison with other methods. In addition, indicators for multidimensional prediction of agricultural product productivity were developed in Uzbekistan.