Enhancing slope failure forecasting model by implementing Archimedean copula to model the error-term

Purpose. In the open-pit mining, monitoring slope failure is a crucial activity. Many mining projects used a manual approach to measure the slope displacement, but an advanced technology (e.g., slope stability radar) offers more precise data collection. Despite these technological strides, accurately predicting slope failure remains essential to prevent costly incidents and interrupted production. Simple linear regression is commonly and widely used with time on the x-axis and inverse-velocity on y-axis, is a prevalent method. However, it often fails to forecast slope failure accurately, as the events tend to occur sooner than predicted. This misprediction might be attributed to the oversight of error-term within the model. Methods. The error-term distinct structure was analyzed to improve the accuracy of the existing linear regression model. To address this, copula models were employed, as they effectively capture the complex dependence patterns among random variables, providing a robust method for incorporating the error-term analysis into the model. Findings. The proposed approach showed its powerful technique to predict slope failure more accurately than the existing simple linear regression model. According to the slope failure datasets, the linear regression model was y = 50475.270 – 1.123 x. Furthermore, the error-term was modeled through Reversed Gumbel-Hougaard (GH-RR) copula model under parameter θ = 1.12. As a result, the prediction missed by 86.4 seconds, compared to 277.284 seconds when using the existing approach (i.e., linear regression without copula-based error-term modeling). Originality. This approach emphasizes the analysis and incorporation of the error-term in the model, which is often overlooked in the simple linear regression method commonly used. Practical implications. Implementing the error-term copula-based model could significantly improve the accuracy of slope failure predictions, thereby preventing costly incidents and ensuring uninterrupted production in mining projects.