Application of Various Machine Learning Techniques in Predicting Total Organic Carbon from Well Logs

Table 2

Summary of different research studies that employed AI techniques to predict the TOC.

Ref	Data sources	Data points	AI tools	Inputs	Accuracy (R^2)
[8]	Barnett and Devonian shale	442	ANN	RHOB, GR, Δt, FR	0.89–0.93
[9]	Shahejie formation	125	CNN	CNP, RHOB, GR, Δt, FR	0.83
[14]	Kazhdumi formation	31	FL	CNP, RHOB, GR, Δt, FR, K, Th, Ur	0.94
[17]	Devonian and Barnett shales	+500	FNN, SVM	RHOB, GR, Δt, FR	0.74–0.77
[18]	Barnett shale	645	FL	RHOB, GR, Δt, FR	0.91
[25]	Jiumenchong formation	31	SVM	CNP, RHOB, GR, Δt, FR, K, Th, Ur	0.69
[27]	Kangan-Dalan formation	124	FL, NF, NN	CNP, RHOB, GR, Δt, FR	0.85
[28]	Kazhdomi and Kangan-Dalan formations	78	ANN	GR, Δt, FR, K, Th	0.89
[29]	Ordos basin and Canning basin	NA	GPR	CNP, RHOB, GR, Δt, FR, K, Th, Ur	NA
[30]	Gadvan formation	2875	ANN, FL	CNP, RHOB, Δt, FR	0.78–0.99
[31]	Kazhdomi and Kangan-Dalan formations	200	ANN	CNP, GR, Δt, FR, K, Th	NA
[32]	NA	70	ANN	Δt, FR	0.98
[33]	Khatatba and Ras Qattara formations	54	ANN	CNP, RHOB, GR, Δt, FR	0.96
[34]	Beibu Gulf basin	18	SVM	RHOB, GR, Δt, SP, FR	0.75
[35]	Sichuan basin	185	ELM, ANN	CNP, RHOB, GR, Δt, FR, K, Th, Ur	0.87–0.91
[36]	Tonghua basin	215	ANN, SVM	CNP, RHOB, GR, Δt, SP, FR, K, Th, Ur	0.9–0.93
[37]	Barnett and Duvernay shales	460	ANN	RHOB, GR, Δt, FR	0.98
[38]	Barnett shale	442	ANN	RHOB, GR, Δt, FR	0.93
[39]	Barnett shale	+800	ANFIS, FNN, SVM	RHOB, GR, Δt, FR	0.82–0.87
[40]	Qaidam basin	19	ANN	RHOB, GR, Δt, FR	NA
[41]	Bohai Bay, Sichuan, Ordos and western Canada sedimentary basins	353	ANN, SVM	RHOB, GR, Δt, FR	0.89