Analyzing Customer Psychological and Behavioral Attributes in Corporate Social Responsibility in Supply Chain Management Using a Multilayer Perceptron Approach

Despite the fact that CSR is currently important in the retail industry, academic research on its influence is sparse. Retailers have a great interest in understanding the influence of customer opinion on CSR initiatives because the system serves as a conduit between producers and consumers. Understanding the impact of CSR projects on store bottom lines is critical for contextualising the system's results. The suggested technique includes preprocessing, feature selection, and model training. This step involved cleaning the dataset with a text-processing method that assigned each document three numerical values: document length, token number, and token length. The best approach is to model feature selection using Latent Dirichlet Allocation (LDA). The MLP was used to train the model in the specified approach. The suggested model has an average accuracy rate of 94.35%, which is greater than cutting-edge alternatives like as SVM and CNN.