Science & Engineering Research Support soCiety (SERSC)
Identifying structure in high dimensional data spaces is a difficult problem. In this paper the authors illustrate a new method of visualizing and projecting time series data using reservoir computing with clustering algorithms. They show the advantages of using clustering with reservoir to visualize data. Then they extend the clustering algorithm and use a fixed latent space to preserve the topology in the projection. They illustrate the method using airport and financial time series data.