A Hybrid Autoencoder with Temporal Fusion Transformer Model for Accurate Dust Prediction in Mining Environments

Abstract

This study presents a systematic framework for data analysis and dust prediction using Machine Learning (ML) techniques. The Air Experimental Dataset is used to collect input data, which is then preprocessed using advanced missing value handling and data cleaning to increase the quality and dependability of the data. Meaningful and structured features are then extracted from the dataset using feature engineering techniques like data visualisation and normalisation. To capture intricate patterns and temporal relationships, a Hybrid Autoencoder with Temporal Fusion Transformer (TFT) is used in the model selection stage using the processed data. The metrics are used to assess the model’s performance. According to experimental results, the proposed model achieves 98% accuracy, precision, recall, and F1-score, indicating robust and consistent prediction performance.