Classifying rockburst with confidence:A novel conformal prediction approach

The scientific community recognizes the seriousness of rockbursts and the need for effective mitigation measures.The literature reports various successful applications of machine learning(ML)models for rockburst assessment;however,a significant question remains unanswered:How reliable are these models,and at what confidence level are classifications made?Typically,ML models output single rockburst grade even in the face of intricate and out-of-distribution samples,without any associated confidence value.Given the susceptibility of ML models to errors,it becomes imperative to quantify their uncertainty to prevent consequential failures.To address this issue,we propose a conformal prediction(CP)framework built on traditional ML models(extreme gradient boosting and random forest)to generate valid classifications of rockburst while producing a measure of confidence for its output.The proposed framework guarantees marginal coverage and,in most cases,conditional coverage on the test dataset.The CP was evaluated on a rockburst case in the Sanshandao Gold Mine in China,where it achieved high coverage and efficiency at applicable confidence levels.Significantly,the CP identified several“confident”classifications from the traditional ML model as unreliable,necessitating expert verification for informed decision-making.The proposed framework improves the reliability and accuracy of rockburst assessments,with the potential to bolster user confidence.

Unconfined compressive strength of MICP and EICP treated sands subjected to cycles of wetting-drying,freezing-thawing and elevated temperature:Experimental and EPR modelling

Microbial-induced carbonate precipitation(MICP)and enzyme-induced carbonate precipitation(EICP)are two bio-cementation techniques,which are relatively new methods of ground improvement.While both techniques share some similarities,they can exhibit different overall behaviours due to the differences in urease enzyme sources and treatment methods.This paper presented 40 unconfined compressive strength(UCS)tests of MICP and EICP treated sand specimens with similar average calcium carbonate(CaCO3)content subjected to cycles of wetting-drying(WD),freezing-thawing(FT)and elevated temperature(fire resistance test e FR and thermogravimetric analysis e TG).The average CaCO3 content after a certain number of WD or FT cycles(ACn)and their corresponding UCS(qn)reduced while the mass loss increased.The EICP treated sand specimens appeared to exhibit a lower resistance to WD and FT cycles than MICP treated specimens possibly due to the presence of unbonded or loosely bonded CaCO3 within the soil matrix,which was subsequently removed during the wetting(during WD)or thawing(during FT)process.FR test and TG analysis showed a significant loss of mass and reduction in CaCO3 content with increased temperatures,possibly due to the thermal decomposition of CaCO3.A complete deterioration of the MICP and EICP treated sand specimens was observed for temperatures above 600C.The observed behaviours are complex and theoretical understanding is far behind to develop a constitutive model to predict qn.Therefore,a multi-objective evolutionary genetic algorithm(GA)that deals with pseudo-polynomial structures,known as evolutionary polynomial regression(EPR),was used to seek three choices from millions of polynomial models.The best EPR model produced an excellent prediction of qn with a minimum sum of squares error(SSE)of 2.392,mean squared error(MSE)of 0.075,root mean square error(RMSE)of 0.273 and a maximum coefficient of determination of 0.939.