@article{scholars13794,
            year = {2020},
             doi = {10.32802/ASMSCJ.2020.434},
          volume = {13},
            note = {cited By 3},
         journal = {ASM Science Journal},
       publisher = {Akademi Sains Malaysia},
           title = {Correlation model development for saybolt colour of condensates and light crude oils},
          author = {Khor, C. S. and Sofia Nurazrin, N. N. and Hanafi, F. M. and Asallehan, F. N. and Rosman, N. Z. and Leam, J. J. and Dass, S. C. and Zainal Abidin, S. A. and Anuar, F. S.},
            issn = {18236782},
             url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85089344913&doi=10.32802\%2fASMSCJ.2020.434&partnerID=40&md5=185402307c4dae5fa0efd059be6c58fa},
        abstract = {Saybolt colour or number is a measured physical property of petroleum condensates and light crude oils which can be used as a quality indicator. As an alternative approach to the laboratory-based colour measurement method, this work aims to determine the influential physical properties in predicting Saybolt colour by applying a regression modelling approach. Data available on Saybolt colour and several physical properties are obtained from assay reports for condensates and light crude oils of Malaysian oil and gas fields. Other unavailable but potentially influential properties are estimated using a commercial process simulation software, iCON. The properties identified as explanatory variables in this study are refractive index, kinematic viscosity at 40C, and characterization factor. This machine learning problem gives rise to applying multiple linear regression techniques based on a backward elimination approach in developing a correlation to predict Saybolt colour using the identified key properties of characterization factor, kinematic viscosity at 40C, and refractive index. {\^A}{\copyright} 2020 Akademi Sains Malaysia.}
}