Ensemble Learning for Predicting TOC from Well-Logs of the Unconventional Goldwyer Shale

Precise estimation of total organic carbon (TOC) is extremely important for the successful characterization an unconventional shale reservoir. Indirect traditional continuous TOC prediction methods from well-logs fail to provide accurate in complex and heterogeneous reservoirs. A workflow proposed predict a profile through various ensemble learning regression models Goldwyer formation Canning Basin, WA. 283 data points ten wells available Rock-Eval analysis core specimen where each sample point contains three five petrophysical logs. The varies largely, ranging 0.16 wt % 4.47 with average 1.20 %. In addition conventional MLR method, four supervised machine methods, i.e., ANN, RF, SVM, GB are trained, validated, tested using approach. To ensure robust prediction, aggregated model predictor designed by combining ensemble-based models. achieved accuracy R2 value 87%. Careful preparation feature selection, reconstruction corrupted or missing logs, implementation optimization have improved significantly compared single