Improvement of K-means Clustering Using Patents Metadata

摘要

Over time, many clustering methods were proposed, but there are many specific areas where adaptations, customizations and modifications of classical clustering algorithms are needed in order to achieve better results. The present article proposes a technique which uses a custom patent model, aiming to improve the quality of clustering by emphasizing the importance of various patent metadata. This can be achieved by computing different weights for different patent metadata attributes, which are considered to be valuable information.