Prediction of bubble pressure using machine learning

• Autor: Gil, Oscar
• Formato: artículo
• Fecha de Publicación: 2023
• Institución: Universidad La Salle
• Repositorio: Revistas - Universidad La Salle
• Lenguaje: español
• OAI Identifier: oai:ojs.revistas.ulasalle.edu.pe:article/82
• Enlace del recurso: https://revistas.ulasalle.edu.pe/innosoft/article/view/82; https://doi.org/10.48168/innosoft.s11.a82; https://purl.org/42411/s11/a82; https://n2t.net/ark:/42411/s11/a82
• Nivel de acceso: acceso abierto
• Materia: Algorithms; Machine learning; Test method; Bubble pressure; Weka; Algoritmos; Aprendizaje automático; Método de prueba; Presión de burbujeo

In the present study, the collection of machine learning algorithms of the Weka program was used to predict the bubble pressure of 36 oil samples, determining the accuracy of their results with the 10-fold cross-validation test method. Subsequently, for comparison purposes, the bubble pressures were calculated with the correlation generated in the work from which the samples were taken and their results were more precise than those obtained by the algorithms in 4 of the 7 performance metrics used. Due to this situation, and considering that the correlation was evaluated with the same data with which it was generated, the test method was changed to validation with the training data and the bubble pressures were predicted again. Other things being equal, machine learning was more accurate than correlation on all performance metrics.