Correlation and root cause analysis of trace data using an unsupervised autoencoder

A computer program product for determining correlations in software application trace data, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

• receive a selection of a parameter of a plurality of parameters represented in an autoencoder;
• in response, identify one or more hidden layer nodes in the autoencoder connected to an output node in the autoencoder corresponding to the selected parameter;
• identify one or more other output nodes in the autoencoder connected to the one or more hidden layer nodes;
• retrieve weights assigned to edges between the one or more hidden layer nodes and the one or more other output nodes;
• integrate a viewer tool with the autoencoder, where the plurality of parameters represented in the autoencoder are displayed on a user interface;
• calculate, by the analyzer system, correlation values between the output node corresponding to the selected parameter and each of the other output nodes,
• wherein one or more other parameters correlated with the selected parameter are graphically highlighted to represent a strength of the correlation, in response to a user clicking on an error message within the viewer tool;
• and
• output the correlation values.

A system comprising:

• a processor;
• and
• a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:
• receive a selection of a parameter of a plurality of parameters represented in an autoencoder;
• in response, identify one or more hidden layer nodes in the autoencoder connected to an output node in the autoencoder corresponding to the selected parameter;
• identify one or more other output nodes in the autoencoder connected to the one or more hidden layer nodes;
• retrieve weights assigned to edges between the one or more hidden layer nodes and the one or more other output nodes;
• integrate a viewer tool with the autoencoder, where the plurality of parameters represented in the autoencoder are displayed on a user interface;
• calculate, by the analyzer system, correlation values between the output node corresponding to the selected parameter and each of the other output nodes,
• wherein one or more other parameters correlated with the selected parameter are graphically highlighted to represent a strength of the correlation, in response to a user clicking on an error message within the viewer tool;
• and
• output the correlation values.