US 12,346,237 B2
Crash bug component prediction system to identify components as potential bug sources
Yang Xu, Xi'an (CN); Yong Li, Xi'an (CN); Hyun Deok Choi, Seoul (KR); Qiao-Luan Xie, Xi'an (CN); and Chao Liu, Xi'an (CN)
Assigned to SAP SE, Walldorf (DE)
Filed by SAP SE, Walldorf (DE)
Filed on Mar. 6, 2023, as Appl. No. 18/178,618.
Prior Publication US 2024/0303179 A1, Sep. 12, 2024
Int. Cl. G06F 8/70 (2018.01); G06F 8/77 (2018.01); G06F 11/362 (2025.01); G06F 11/3668 (2025.01); G06N 5/022 (2023.01)
CPC G06F 11/3636 (2013.01) [G06F 8/77 (2013.01); G06F 11/3692 (2013.01); G06N 5/022 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method for identifying components in crash reports generated in response to crashes of software systems, the method being executed by one or more processors and comprising:
receiving a crash report provided as a computer-readable file;
providing a stack trace from the crash report;
adding component information to the stack trace to provide a stack trace and component table comprising a component column populated with a set of the components, a component position column populated with a set of component positions, each component in the set of components being associated with a respective component position in the set of component positions, a function name column populated with a set of function names, each component in the set of the components being associated with a sub-set of function names, and a function value column populated with a set of function values, each component being associated with a sub-set of function values;
for each component identified in the stack trace, determining a set of features;
processing sets of features through a machine learning (ML) model to provide a prediction identifying a component of the set of the components as a bug component, the ML model having been trained using training data comprising, for each stack trace in a set of historical stack traces, a set of historical features and a label, the set of historical features corresponding to a bug component for the respective stack trace; and
assigning the bug component for resolution through a crash management system.