On Optimizing Gaussian Function based Similarity Metric in Computational Biology


Gaussian kernel is a popular kernel function often used in various computational biology and their applications to measure the similarity between two samples in a dataset. It can be used not only in the unsupervised situation but also in the supervised case. However, the scalar parameter belta () in Gaussian kernel function as shown in the following equation significantly affects the final results. This online data-driven tool is used to calculate the optimal belta () parameter based on the given dataset.

When using this online service, the user is required to submit their dataset with a file named *.txt, for example "in.txt" .
(2) For very large-size or very high-dimensional datasets, the users need to download the stand alone program.


For Supervised Case:

Please upload your dataset (Example): 



For Unsupervised Case:

Please upload your dataset (Example):


Jian-Bo Lei, Jiang-Bo Yin, and Hong-Bin Shen, GFO: A data driven approach for optimizing Gaussian function based similarity metric in computational biology, Neurocomputing, 2013, 99: 307-315.

Jiang-Bo Yin, Tao Li, and Hong-Bin Shen, Gaussian kernel optimization: complex problem and a simple solution, Neurocomputing, 2011, 74: 3816-3822.

Contact @ Hong-Bin