Microsoft patents system that groups support tickets by fix commands

1. A computer-implemented method comprising: extracting, by a network system, commands used to resolve technical problems indicated in a plurality of resolved support tickets, a technical problem comprising an issue experienced in a technical system, each of the commands comprising at least one instruction that causes the network system to perform one or more operations; creating, by one or more hardware processors of the network system, clusters of resolved support tickets based on similarity of the at least one instruction associated with each of the commands used to resolve technical problems that have been extracted from the plurality of resolved support tickets and not based on problem statements in the resolved support tickets, the problem statements each indicating the technical problem, the clusters further based on a number of words in the at least one instruction associated with each of the commands; for each cluster of the resolved support tickets, extracting, by the network system, the problem statements from the resolved support tickets in the same cluster; training, by the network system, a machine learning model with training data comprising the extracted problem statements from the resolved tickets that have been clustered together based on common commands, each of the common commands comprising at least one common instruction that causes the network system to perform one or more common operations, a cluster number for each cluster and a vector representing the extracted problem statements for each cluster; maintaining the machine learning model for use during runtime to resolve technical problems by applying the machine learning model to a problem statement extracted from a new support ticket; applying the common commands that correspond to the technical problems based on clusters identified by the machine learning model; identifying feedback indicative of whether the common commands successfully resolved the technical problem; and based on the feedback, automatically retraining the machine learning model by changing a number of the clusters.