Solving Agricultural Price Recommendation Problem Using Smart Reading Algorithms

Abstract

An optimization problem with time series forecasting constraints (OPTSFC) is defined as one with constraints given as prediction results from time series. OPTSFC can be solved in stages using prediction and optimization algorithms. In this study, the OPTSFC problem was solved using a smart reading algorithm (SRA), a strategy that uses various alternative settlement techniques according to the available criteria. The techniques used in the SRA-OPTSFC include descriptive statistics, Regression (Linear), recurrent neural network, and particle swarm optimization. In this study, the OPTSFC model and the SRA-OPTSFC algorithm are implemented in the agricultural price recommendation problem. The results of the implementation show that different criteria produce different optimal solutions. A high prediction accuracy value produces an optimal solution close to the prediction results. This has been proved in the predicted value theorem, as illustrated through research experiments. Furthermore, the experimental results show the importance of the time series changes from low to high categorization accompanied by changes in the accuracy value to 93.6%, indicating a narrowing of the feasible area by 99.65% toward the time series prediction results.