Studies on Problems of data clustering are continued

At the Institute of Information Technology of ANAS, the next scientific seminar of the Department №13 dedicated to the discusssion of  " Data Clustering Algorithms and Applications " book was held.

The magistrate of the Institute, Shelale Mansurova, spoke about the intellectual analysis of the data covering the second chapter of the book, the selection of signs as one of the solutions to high-dimensional problems in machine learning and the clustering of data.

Reporter gave information about key stages and models for selecting traits as one of the solutions to high-dimensional problem-solving issues in intellectual analysis. She stated that there are key stages for selecting symptoms such as "subset generation", "subset assessment", "stopping criterion", and "result validation". Basically, the "filter", "wrapper", "hybrid" models are used to select signals.

She mentioned basic methods used in the cluster analysis of data, including "partitioning", "hierarchical", "density-based" methods. Also,  algorithms for general data, text data, streaming data, related data  and so on.  were discussed and alghoritmssuch as Spectral FS, Laplacian Score (LS), MultiClaster FS (MCFS), Feature weighting k-Means were noted. Some open problems such as Model selection, scaling and so on were touched upon.  

In the end, the views were exchanged, questions were answered. Doctor of technical sciences Ramiz Aliguliyev gave his recommendations on further deepening of research.

© All rights reserved. Citing to www.ict.az is necessary upon using news