US 11,817,994 B2
Time series trend root cause identification
Jifu Zhao, Urbana, IL (US); Kevin Andrew Perkins, Champaign, IL (US); Mithilesh Nanjamanaidu Srinivasan Rangavadivel, Champaign, IL (US); and Matthew Robert Ahrens, Champaign, IL (US)
Assigned to YAHOO ASSETS LLC, New York, NY (US)
Filed by Verizon Media Inc., New York, NY (US)
Filed on Jan. 25, 2021, as Appl. No. 17/156,990.
Prior Publication US 2022/0239549 A1, Jul. 28, 2022
Int. Cl. H04L 41/0631 (2022.01); H04L 41/069 (2022.01); H04L 43/04 (2022.01); H04L 43/067 (2022.01); H04L 41/0604 (2022.01)
CPC H04L 41/064 (2013.01) [H04L 41/0604 (2013.01); H04L 41/069 (2013.01); H04L 43/04 (2013.01); H04L 43/067 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
executing, on a processor of a computing device, instructions that cause the computing device to perform operations, the operations comprising:
identifying a set of data, wherein the set of data is organized into multi-dimensional time series data comprising one or more dimensions of measured elements comprising a first measured element and a second measured element;
identifying an overall trend of the multi-dimensional time series data;
identifying element trends, for the measured elements, comprising a first element trend of the first measured element and a second element trend of the second measured element;
calculating weighted correlations between the element trends of the measured elements and the overall trend,
wherein the weighted correlations comprise:
a first weighted correlation between (i) the first element trend and (ii) the overall trend; and
a second weighted correlation between (i) the second element trend and (ii) the overall trend;
determining aggregate weighted correlations of measured element combinations comprising an aggregate weighted correlation for a combination of (i) the first measured element having the first element trend and (ii) the second measured element having the second element trend;
using the overall trend to evaluate the weighted correlations of the measured elements and the aggregate weighted correlations of the measured element combinations to iteratively identify a set of measured elements having a threshold correlation to the overall trend, wherein the using the overall trend to evaluate the weighted correlations and the aggregate weighted correlations comprises including, in the set of measured elements, the combination of (i) the first measured element having the first element trend and (ii) the second measured element having the second element trend based upon the aggregate weighted correlation for the combination having the threshold correlation to the overall trend;
providing an indication of the set of measured elements, including the combination of (i) the first measured element having the first element trend and (ii) the second measured element having the second element trend, as a root cause of the overall trend, wherein the indication includes a number of top sorted root causes, comprising the root cause, populated in a first interface displayed through a user interface;
transitioning the user interface from displaying the first interface to displaying a second interface populated with information related to a dimension of at least one of the first measured element or the second measured element;
evaluating the indication of the set of measured elements to determine whether the root cause of the overall trend corresponds to an operational issue or a non-operational causation factor;
responsive to determining that the root cause of the overall trend corresponds to the operational issue, generating a resolution request for the operational issue corresponding to the root cause of the overall trend; and
routing the resolution request to an entity for resolution.