Paper Title
Using Data Mining Technology to Analyze Traditional Gold Mining in Sudan
Abstract
The traditional gold mining sector is spreading throughout most parts of Sudan, where more than two million
people are working in it. Producing about 80% of the total amount of mined gold in the country. This study used the
methodology of extracting the data to assist in making decisions using five models support vector machine (SVM), logistic
regression (LR), naive Bayes (NB),decision tree (DT) and k-nearest neighbors (K-NN) and a fair comparison was made
between their performance. These models classify the state of companies into two categories: operating, companies with
high productivity and suspended with poor productivity. The learning process took place in four stages: initial data
processing, training, testing, and verification. The results showed that the proposed models performed the classification task
to reveal the state of the company's work. The models achieved a high level of accuracy for the DT, LR, and NB classifier
1.00, 0.91 and 0.81. The SVM and K-NN models decreased by 0.57 and 0.53 compared to the other models.
Keywords - Decision Tree; Support Vector Machine; K-nearest neighbor; Naive Bayes; Logistic Regression; Traditional
Mining.