UCLA

The spectral clustering algorithm is an algorithm for putting N data points in an I-dimensional space into several clusters. Each cluster is parameterized by its similarity, which means that the points in the same group are similar and points in different groups are dissimilar to each other. Recently, spectral clustering has become an increasingly adopted tool and has been applied in many areas such as statistics, machine learning, pattern recognition, data mining, and image processing. This paper discusses these methods in details and later on introduce the relationship between spectral clustering and k-means clustering and spectral clustering's applications in image segmentation, educational data mining, entity resolution and speech separation.