<![CDATA[The development of the world economy is currently undergoing a very complex phase characterised by very different changes. These changes have a significant impact on the cost of living and quality of life of people in various countries around the world. Factors such as government policies, economy, income, transport, fuel play an important role in various changes in the world economy. Then this research aims to cluster the cost of living on the "Global Living Cost" dataset against socioeconomic status and also determine the performance of the DBSCAN algorithm on several types of "Global Living Cost" datasets that have been modified by perturbing using jitter position, combination of jitter position and scale, and noise perturbation then also using several dimensionality reduction techniques such as PCA, UMAP, t-SNE, and ICA. The data used is taken from "Kaggle. com", then obtained similarity results using the jaccard similarity method that the combination of clustering and dimensionality reduction algorithms for the "Global Living Cost" dataset from the best is DBSCAN and ICA with the highest similarity score of several types of datasets, namely 1.0, DBSCAN and PCA with the highest similarity score of 0.998769735493131, DBSCAN and t-SNE with the highest similarity score of 0.9995897435897436, and the last is DBSCAN and UMAP with the highest similarity score of 0.8065233506300964. The conclusion obtained is that the DBSCAN algorithm is able to work very well for different types of datasets.]]>
Copyrights © 2024
