Method to monitor accuracy of analytics in a mobile communication system
An apparatus, in a model training related network element, is provided, the apparatus comprising: at least one processor and at least one memory including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus at least to perform: receiving a monitoring correlation identifier, subscribing to analytics related information provided by an analytics network element from a storage network element based on the received monitoring correlation identifier, retrieving the analytics related information provided by the analytics network element from the storage network element, and calculating the accuracy of the analytics based on the retrieved analytics related information.
1 . An apparatus comprising:
at least one processor; and
at least one memory storing computer program code for a network data analytics function (NWDAF) containing machine learning model training logical function (MTLF), wherein the computer program code, when executed by the at least one processor, cause the apparatus at least to perform:
receiving, from a NWDAF containing an analytics logical function (AnLF), a request for a machine learning (ML) model, the request comprising a monitoring correlation identifier;
retrieving, from an analytics data repository function (ADRF) using the monitoring correlation identifier, analytics related information stored in the ADRF in association with the monitoring correlation identifier, the analytics related information including predictions provided by the ML model and measured data observed at times the predictions were made that are stored in the ADRF in association with the monitoring correlation identifier; and
computing an accuracy of the predictions provided by the ML model comprised in the analytics related information that is retrieved, wherein an accuracy of the ML model is computed using the predictions and the measured data observed at the time the predictions were been made;
determining, based on the accuracy of the predictions provided by the ML model, whether the ML model is to be re-trained; and
re-training the ML model based on determining that the ML model is to be re-trained.
2 . The apparatus as claimed in claim 1 , wherein the computer program code, when executed by the at least one processor, further causes the apparatus to perform:
providing the re-trained ML model to the NWDAF containing the AnLF.
3 . The apparatus as claimed in claim 1 , wherein the computer program code, when executed by the at least one processor, further causes the apparatus to perform:
creating an accuracy report based on or comprising the accuracy of predictions; and
providing the accuracy report to the NWDAF containing the AnLF.
4 . The apparatus as claimed in claim 1 , wherein the computer program code, when executed by the at least one processor, further causes the apparatus to perform:
providing, to the NWDAF containing the AnLF, the ML model.