Technique for setting a decision threshold of a communication network analytics system

1 . A method of setting one or more decision thresholds of a network analytics system, wherein the analytics system is configured to analyze statistics pertaining to a service quality-related, SQR, parameter acquired for a communication network and statistics pertaining to a root cause-related, RCR, parameter acquired for the communication network, wherein the SQR parameter correlates with the RCR parameter, and wherein a first parameter is defined to be one of the SQR parameter and the RCR parameter and a second parameter is defined to be the other of the SQR parameter and the RCR parameter, the method comprising:

obtaining first statistics for the first parameter in situations in which the second parameter assumes critical second parameter values;

obtaining second statistics for the first parameter in situations in which the second parameter assumes non-critical second parameter values; and

setting a first decision threshold of the analytics system based on an analysis of the first statistics and the second statistics, wherein the first decision threshold is defined to separate critical from non-critical first parameter values,

wherein the critical second parameter values are separated from the non-critical second parameter values by one of a second decision threshold and a binary decision;

wherein for setting the first decision threshold, the first statistics and the second statistics are analyzed to derive at least one of the following two attribute sets:

false critical; false non-critical and

true non-critical; true critical, with

a primary attribute critical or non-critical being associated with first decisions for the first parameter values, and

a secondary attribute true or false being associated with second decisions for the second parameter values dependent on whether or not the respectively associated first decision can be confirmed;

wherein the first decision threshold is set to configure at least one of

a. relative probabilities or a ratio of false critical; false non-critical decisions; and

b. relative probabilities or a ratio of true critical; true non-critical decisions; and

wherein the first decision threshold is set to equalize, based on the first statistics and the second statistics, one of

a. the probabilities or numbers of false critical and false non-critical decisions; and

b. the probabilities or numbers of true critical and true non-critical decisions.

2 . The method of claim 1 , wherein

at least one of the first statistics and the second statistics takes the form of a parameter value distribution,

at least one of the RCR parameter and the SQR parameter assumes continuous parameter values, and

an associated parameter value distribution is obtained by applying a discretization operation to the continuous parameter values.

3 . The method of claim 1 , wherein

at least one of the RCR parameter and the SQR parameter is a key performance indicator, KPI, pertaining to the communication network.

4 . The method of claim 1 , wherein

the SQR parameter is derived from, or indicative of, at least one of one or more events or counters relating to a user plane of a core domain of the communication network; and

the RCR parameter is derived from, or indicative of, at least one of:

a. one or more radio environment measurements;

b. one or radio signalling events; and

c. one or more events or counters relating to a core domain of the communication network.

5 . The method of claim 1 , wherein

the SQR parameter correlates with the RCR parameter in that a possible root cause as indicated by a critical RCR parameter value potentially leads to a service quality degradation as indicated by a critical SQR parameter value.

6 . The method of claim 1 , wherein

the first parameter is the SQR parameter and the second parameter is the RCR parameter.

7 . The method of claim 6 , wherein

the first statistics are obtained for a dedicated service out of a set of services supported by the communication network; and

a dedicated first threshold is set for each of two or more dedicated services out of the set of services.

8 . The method of claim 1 , further comprising:

obtaining third statistics for the second parameter in situations in which the first parameter assumes critical first parameter values;

obtaining fourth statistics for the second parameter in situations in which the first parameter assumes non-critical first parameter values; and

setting the second decision threshold based on the third statistics and the fourth statistics.

9 . The method of claim 8 , wherein

for setting the second decision threshold, the third statistics and the fourth statistics are analyzed to derive at least one of the following two attribute sets:

false critical; false non-critical and

true non-critical; true critical, with

a primary attribute critical or non-critical being associated with second decisions for the second parameter values, and

a secondary attribute true or false being associated with first decisions for the first parameter values dependent on whether or not the respectively associated second decision can be confirmed.

10 . The method of claim 9 , wherein

the second decision threshold is set to configure at least one of

a. relative probabilities or a ratio of false critical; false non-critical decisions; and

b. relative probabilities or a ratio of true critical; true non-critical decisions.

11 . The method of claim 10 , wherein

the second decision threshold is set to equalize, based on the first statistics and the second statistics, one of

a. the probabilities or numbers of false critical and false non-critical decisions; and

b. the probabilities or numbers of true critical and true non-critical decisions.

12 . The method of claim 8 , wherein

setting of the first decision threshold and setting of the second decision threshold are iteratively repeated.

13 . The method of claim 1 , comprising

detecting, based on one or more critical values of the SQR parameter, a service quality degradation;

performing, in response to detection of the service quality degradation, a root cause analysis based on one or more of the RCR parameter values; and

determining, based on one or more critical values of the RCR parameter, a possible root cause.

14 . The method of claim 1 , wherein

the analytics system is rule-based, and wherein the first decision threshold belongs to a first network analytics rule configured to detect the service quality degradation.

15 . A non-transitory computer-readable medium storing instructions thereon that when executed by one or more processors cause a computer to perform the method of claim 1 .

16 . The method of claim 1 , wherein the first decision threshold is set to equalize, based on the first statistics and the second statistics, the probabilities of false critical and false non-critical decisions.

17 . The method of claim 1 , wherein the first decision threshold is set to equalize, based on the first statistics and the second statistics, the numbers of false critical and false non-critical decisions.

18 . The method of claim 1 , wherein the first decision threshold is set to equalize, based on the first statistics and the second statistics, the probabilities of true critical and true non-critical decisions.

19 . The method of claim 1 , wherein the first decision threshold is set to equalize, based on the first statistics and the second statistics, the numbers of true critical and true non-critical decisions.