JASIC Volume. 6, Issue 1 (2025)

  1. Home
  2. Volume. 6, Issue 1

Contributor(s)

Ogunleye Timothy A
 

Keywords

Kerogen Classification Machine Learning Algorithms Geochemical Analysis Hydrocarbon
 

Download Full-text (PDF)

... Download File [ 0.45 MB ]
 
Go Back

Predictive modelling of Kerogen types using supervised machine learning algorithms: A geochemical study of the Niger Delta Basin, Nigeria

Abstract: Kerogen classification helps hydrocarbon explorers assess the ability of rocks to produce oil and natural gas. The traditional methods of Rock-eval pyrolysis and elemental analysis still rely on previous data interpretations and often possess inaccuracies. We propose using machine learning approaches to enhance the accuracy and speed of kerogen type classification in the Niger Delta Basin utilizing geochemical data. Geochemical properties such as S1, S2, S3, Tmax, HI, OI, TOC, and PI were derived from the analyzed oil and gas well samples. Various supervised machine learning algorithms such as Random Forest, Gradient Boosting, Support Vector Machines (SVM) and Decision Trees were used to assign kerogen to Types I to IV. Performance measures were determined by computing accuracy, precision, recall and the F1 score. Ensemble methods showed the highest levels of precision and reliability among all the algorithms. It was determined that Oxygen Index, S3 and Tmax played the central role in determining kerogen quality compared to other characteristics. The algorithms (Decision Tree, Random Forest, Gradient Boosting, Ada Boosting, Bagging and Extra Trees) showed comparable results in classification precision. Applying machine learning techniques substantially improves the accuracy and objectivity of kerogen classification and exploring ensemble methods produces geoscientific results that are much more precise when compared with those obtained using conventional methods. Subsequently, the author found that only three of the features showed nearly equal percent contributions. The three highest percentage contributions came from oxygen index (17.5%), carbon dioxide generated through pyrolysis (17.3%) and temperature (16.2%). This information has been added to the current knowledge available in the field of geosciences
 
 
 

Journal of Applied Science, Information and Computing (ISSN : 3007-8903), 2025 | All Articles