US 11,928,141 B1
Method, electronic device, and computer program product for retrieving service request
Jiacheng Ni, Shanghai (CN); Zijia Wang, WeiFang (CN); and Jinpeng Liu, Shanghai (CN)
Assigned to Dell Products L.P., Round Rock, TX (US)
Filed by Dell Products L.P., Round Rock, TX (US)
Filed on Nov. 29, 2022, as Appl. No. 18/071,247.
Claims priority of application No. 202211288599.4 (CN), filed on Oct. 20, 2022.
Int. Cl. G06F 16/33 (2019.01); G06F 16/35 (2019.01)
CPC G06F 16/334 (2019.01) [G06F 16/35 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method for retrieving service requests, comprising:
determining, based on an acquired service request, a hash value of the service request;
determining a plurality of request pairs based on a plurality of correlations between the hash value of the service request and a plurality of hash values of a plurality of historical service requests, wherein each of the plurality of request pairs comprises the service request and one historical service request, utilizing a filtering operation based on a first indicator in a first stage of a multi-stage similarity modeling process;
determining a semantic correlation between the service request and the one historical service request in each of the plurality of request pairs;
determining, based on the determined semantic correlation of each request pair, a probability indicating that the service request and the historical service request in the request pair use the same solution, to provide a classification result as a second indicator in a second stage of the multi-stage similarity modeling process; and
determining a retrieved historical service request based on the probability;
wherein labels are associated with respective request pairs in conjunction with the multi-stage similarity modeling process, a given one of the labels having one of a first value to indicate that two requests in the corresponding pair use the same solution and a second value different than the first value to indicate that the two requests in the corresponding pair do not use the same solution;
the classification result being generated by a machine learning model trained utilizing at least a portion of the labeled request pairs.