A comprehensive model of demand prediction based on hybrid artificial intelligence and metaheuristic algorithms: A case study in dairy industry

This paper presents a multi-stage model for accurate prediction of demand for dairy products (DDP) by the use of artificial intelligence tools including Multi-Layer Perceptron (MLP), Adaptive-Neural-based Fuzzy Inference System (ANFIS), and Support Vector Regression (SVR). The innovation of this work is the improvement of artificial intelligence tools with various meta-heuristic algorithms including Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Invasive Weed Optimization (IWO), and Cultural Algorithm (CA). First, the best combination of factors with can affect the DDP is determined by solving a feature selection optimization problem. Then, the artificial intelligent tools are improved with the goal of making a prediction with minimal error. The results indicate that demographic behavior and inflation rate have the greatest impact on dairy consumption in Iran. Moreover, PSO still exhibits a better performance in feature selection in compare of newcomer meta-heuristic algorithms such as IWO and CA. However, IWO shows the best performance in improving the prediction tools by achieving an error of 0.008 and a coefficient of determination of 95%. The final analysis demonstrates the validity and reliability of the results of the proposed model, as it supports the simultaneous analysis and comparison of the outputs of different tools and methods.