Computer-implemented method of modifying an algorithm operating on a computing system
This invention provides a computer-implemented method of modifying an algorithm operating on a computing system, and a device for implementing said method, the method comprising the steps of: applying the algorithm to a first set of inputs; determining a relevance score for a first input of the first set of inputs based on: a first effectiveness value of the first input, wherein the first effectiveness value represents a contribution of first input to the algorithm, and a first computational cost of the first input, wherein the 1 first computational cost represents the computational resources of using the first input in the algorithm; defining a second set of inputs based on the determined relevance score of the first input; and applying the algorithm to the second set of inputs.
1. A computer-implemented method of modifying an algorithm operating on a computing system, the method comprising the steps of:
applying the algorithm to a first set of inputs;
determining a relevance score for a first input of the first set of inputs based on:
a first effectiveness value of the first input, wherein the first effectiveness value represents a degree of influence of the first input on an output of the algorithm, and
a first computational cost of the first input, wherein the first computational cost represents the computational resources of using the first input in the algorithm;
comparing the relevance score for the first input to an operator defined threshold;
defining a second set of inputs based on the comparison of the relevance score of the first input to the threshold such that:
if the relevance score of the first input does not satisfy the threshold, the second set of inputs is based on the first set of inputs with the first input removed, and
if the first input is a candidate input and the relevance score for the first input satisfies the threshold, the second set of inputs includes accepting the first input; and applying the algorithm to the second set of inputs.
2. A method as claimed in claim 1 , wherein the algorithm operates to select an operating profile from a set of operating profiles for use in a Digital Subscriber Line, DSL, system, each operating profile of the set of operating profiles specifying one or more operating parameters.
3. A method as claimed in claim 1 , wherein the algorithm operates to determine whether to automatically or manually restart a service operating on a Digital Subscriber Line, DSL, system.
4. A method as claimed in claim 1 , further comprising the steps of:
determining a relevance score for a second input of the first set of inputs, wherein the second input is a candidate input, based on:
a second effectiveness value of the second input, wherein the second effectiveness value represents how influential the second input is on the output of the algorithm, and
a second computational cost of the second input, wherein the second computational cost represents the computational resources of using the second input in the algorithm;
comparing the relevance score for the second input to the threshold,
wherein the step of defining the second set of inputs is further based on the comparison of the relevance score for the second input to the threshold such that if the relevance score of the first input does not satisfy the threshold and the relevance score of the second input does satisfy the threshold then the second set of inputs is based on the first set of inputs with the first input removed and further includes accepting the second input.
5. A method as claimed in claim 1 , wherein the second set of inputs includes a new candidate input.
6. A method as claimed in claim 5 , wherein the second computational cost includes one or more of the processing cost, storage cost and transmission cost associated with the second input.
7. A method as claimed in claim 1 , wherein the first computational cost includes one or more of the processing cost, storage cost and transmission cost associated with the first input.
8. A method as claimed in claim 1 , wherein the algorithm is updated by a machine learning process.
9. A method as claimed in claim 1 , wherein the first input represents a combination of parameters.
10. A method as claimed in claim 1 , wherein a value of the first input is determined by a data gathering node, and the first input is pre-processed by the data gathering node.
11. A non-transitory computer-storage medium storing instructions which, when the program is executed by a computer, cause the computer to carry out the method of claim 1 .
12. A device for modifying an algorithm operating on a computing system, the algorithm operating on a first set of inputs, the device comprising a processor configured to:
determine a relevance score for a first input of the first set of inputs based on:
a first effectiveness value of the first input, wherein the first effectiveness value represents a degree of influence of the first input on an output of the algorithm, and
a first computational cost of the first input, wherein the first computational cost represents the computational resources of using the first input in the algorithm;
compare the relevance score for the first input to an operator defined threshold;
define a second set of inputs based on the comparison of the relevance score of the first input to the threshold such that:
if the relevance score of the first input does not satisfy the threshold, the second set of inputs is based on the first set of inputs with the first input removed, and
if the first input is a candidate input and the relevance score for the first input satisfies the threshold, the second set of inputs includes accepting the first input; and
cause the algorithm to use the second set of inputs.
13. The device as claimed in claim 12 , wherein the algorithm is configured to operate to select an operating profile from a set of operating profiles for use in a Digital Subscriber Line (DSL) system, each operating profile of the set of operating profiles specifying one or more operating parameters.
14. The device as claimed in claim 12 , wherein the algorithm is configured to operate to determine whether to automatically or manually restart a service operating on a Digital Subscriber Line (DSL) system.
15. The device as claimed in claim 12 , wherein the device is configured to:
determine a relevance score for a second input of the first set of inputs, wherein the second input is a candidate input, based on:
a second effectiveness value of the second input, wherein the second effectiveness value represents how influential the second input is on the output of the algorithm, and
a second computational cost of the second input, wherein the second computational cost represents the computational resources of using the second input in the algorithm;
compare the relevance score for the second input to the threshold,
wherein the definition of the second set of inputs is further based on the comparison of the relevance score for the second input to the threshold such that if the relevance score of the first input does not satisfy the threshold and the relevance score of the second input does satisfy the threshold then the second set of inputs is based on the first set of inputs with the first input removed and further includes accepting the second input.
16. The device as claimed in claim 12 , wherein the second set of inputs includes a new candidate input.
17. The device as claimed in claim 16 , wherein the second computational cost includes the processing cost, storage cost and/or transmission cost associated with the second input.
18. The device as claimed in claim 12 , wherein the first computational cost includes the processing cost, storage cost and/or transmission cost associated with the first input.
19. The device as claimed in claim 12 , wherein the algorithm is updated by a machine learning process.
20. The device as claimed in claim 12 , wherein a value of the first input is determined by a data gathering node, and the first input is pre-processed by the data gathering node.