Big Data-Driven Analysis of Enterprise Growth Factors Using Neural Networks and Detrended Fluctuation Techniques for Market Prediction and Strategic Planning

This study proposes a novel approach to analyze the social market information for mining enterprises using big data analytics and deep learning. One of the biggest challenges in industrial planning and market prediction is identifying important enterprise development variables from large amounts of unstructured data. To address this issue, we propose the hybrid Deep Neural Network (DNN) model combined with Fully Connected Networks (FCNs) and Convolutional Neural Networks (CNNs) to train the real-world social and market datasets. The model combines the advantages of Noise Data Augmentation (NDA) to improve sensitivity and environmental variability, and Detrended Fluctuation Analysis (DFA) to calculate Hurst components to understand the complexities and consistency of social and market trends. The proposed system effectively classifies the enterprise improvement features in three difficulty levels: Easy, Medium and Hard. The experiments show high classification accuracy even in low-resource edge devices and prove its effectiveness. Overall, this study presents an effective approach for strategic market prediction in resource-constrained settings and also highlights the ability in real-time decision-support systems in mining firms.