Fuzzy Soft Set Based Classification for Rock Dataset

One of the main tasks in geological studies is rock classification. To examine rock samples in this classification usually requires a human expert. Thus, the igneous rocks’ classification task will become challenging because of igneous rocks’ diverse composition. One data mining technique based on Fuzzy soft set can be used for classification. Several similarity measures have been proposed on the fuzzy soft set. In this paper, we conduct an experiment to explore the fuzzy soft set classifier applying several measurement to calculate the similarity, i.e., generalized fuzzy soft sets, similarity based on matching function, similarity based on set theoretic approach, similarity measure based on distance. The classification of igneous rocks is carried out in this experiment based on their chemical composition and compared it in terms of accuracy, precision, and recall. According to our simulation results, the Euclidean distance still outperforms to another measure in terms of classification accuracy, precision, and recall.