CLUSTERING CREDIBILISTE FLOU DES DONNEES PETROLIERES VIA FUZZY C-MEANS.

Our paper deals with the application of the Fuzzy C-Means credibilistic method by adding a one-sided function with a regularisation technique on oil production data. Oil production is important for the development of a country. This sector is facing a lack of data management policy to balance oil exploitation in Onshore and Offshore. The data are grouped into three main clusters and come from the MOTOBA field in the Democratic Republic of Congo, with production from 2018 to 2021 on 10038 records with 5 attributes. According to the results obtained, we have Cluster 1 which groups together wells with high oil production, and which produce a lot of oil but little gas and water. Cluster 2 with wells with balanced production; these wells produce average quantities of oil, gas and water. Cluster 3 with wells that produce a lot of water. Three measures of validity were used: the fuzzy silhouette index with a score of 0.3877, the Sum of Squared Errors (SSE) of 27.5 and the Kullback-Leibler divergence (KL) of 0.3.