Intelligent demand forecasting in marketing sector using concatenated CNN with ANFIS enhanced by heuristic algorithm

Abstract

AbstractThis task introduces a novel demand forecasting method using concatenated Convolutional Neural Network (CNN) with an Adaptive Network-based Fuzzy Inference System (ANFIS). The data regarding the historical demand and sales data in integration with ‘advertising effectiveness, expenditure, promotions, and marketing events data' are collected initially. Then, the first-order statistical metrics and second-order statistical metrics are determined as the significant features of the data. Finally, the forecasting is performed by the concatenation of modified CNN with ANFIS termed Concatenated Learning Model (CLM), in which the CNN learns the optimal features that are forecasted by the ANFIS layer instead of the fully connected layer. Deer Hunting with Modified Wind Angle Search (DH-MWS) is used to enhance the CNN and ANFIS architecture, ensuring better performance during forecasting. Simulation findings demonstrate that when the proposed solution is applied to public data, the store achieves improved accuracies concerning intelligent demand forecasting in the marketing sector.KEYWORDS: Demand forecastingmarketing sectorconcatenated learning modeldeer hunting with modified wind angle search Disclosure statementNo potential conflict of interest was reported by the author(s).Additional informationNotes on contributorsN. Srikanth ReddyN. Srikanth Reddy. A Commerce graduate with Post-Graduation in Management and Doctorate in Management. More than 15 years of experience in education and research. Areas of interest include Marketing, Systems and Analytics.