Parameter set determination for clustering of datasets

• US 10,235,443 B2
• Filed: 03/01/2016
• Issued: 03/19/2019
• Est. Priority Date: 03/01/2016
• Status: Active Grant

• Current Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Original Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Inventors: Hackett-Jones, Emily, Chew, Theam Yong, Kirkby, Stephen, Pallotta, Rob, Enderling, Katherine
• Primary Examiner(s): Mobin, Hasanul
• Application Number:
  US15/057,857
• Publication Number:
  US 20170255688A1
• Time in Patent Office: 1,113 Days
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/006 based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A clustering system selects a parameter set from a plurality of parameter sets associated with a dataset to generate a hyper-optimized cluster of the dataset. Different parameter sets are generated by varying the parameter values based on the Genetic Algorithm or a particle swarming algorithm. A parameter set having a high Fitness score is selected from the different parameter sets and a clustered solution produced using the selected parameter set having a maximum total score is used to produce actionable items.

Claims

1. A clustering cycle optimization system comprising:

a data storage to store parameters for clustering a data set; and

one or more processors to perform operations to;

generate an initial parameter set of a plurality of parameter sets via random selection of parameter values, the initial parameter set including different types of parameters;

cause random changes in values of the different types of parameters in the initial parameter set, wherein the random changes include changes of a predetermined percentage caused in the values of numerical parameters, toggling Boolean parameters and resampling categorical parameters;

generate the plurality of parameter sets based on the random changes

generate a plurality of cluster solutions from each of the plurality of parameter sets, the plurality of cluster solutions generated for each parameter set comprising an optimized cluster solution having a highest total score of the plurality of cluster solutions generated for the parameter set;

obtain a respective fitness score for each parameter set of the plurality of parameter sets, the respective fitness score of the parameter set being based on total scores of the plurality of cluster solutions generated for the parameter set;

select a hyper-optimized parameter set from the plurality of parameter sets based on a comparison of the respective fitness scores of the plurality of parameter sets;

select a hyper-optimized cluster solution from the plurality of cluster solutions generated from the hyper-optimized parameter set, the selection of the hyper-optimized cluster solution being based on comparisons of the total scores of the plurality of cluster solutions generated for the hyper-optimized parameter set; and

perform an actionable item that includes transmission of a communication by an external system to a group of users based on the hyper-optimized cluster solution.

2. The clustering cycle optimization system of claim 1, wherein the predetermined percentage is 5%.

3. The clustering cycle optimization system of claim 1, wherein the system comprises:

a crossover operator that randomly swaps parameters between two of the plurality of parameter sets.

4. The clustering cycle optimization system of claim 1, wherein the hyper-optimized parameter set has a high fitness score relative to fitness scores of other parameter sets of the plurality of parameter sets.

5. The clustering cycle optimization system of claim 4, wherein the hyper-optimized cluster solution has a maximum total score of the plurality of cluster solutions associated with the hyper-optimized parameter set.

6. The clustering cycle optimization system of claim 4, wherein the fitness score of a current parameter set is an average of total scores of upper quartile subset of the plurality of cluster solutions.

7. The clustering cycle optimization system of claim 1, wherein the hyper-optimized cluster solution includes clusters of data that each have attributes for determining the actionable item.

8. A computer implemented method for hyper-optimizing a clustering engine, comprising:

generating an initial parameter set of a plurality of parameter sets via random selection of parameter values, the initial parameter set including different types of parameters;

causing random changes in values of the different types of parameters in the initial parameter set, wherein the random changes include changes of a predetermined percentage caused in the values of numerical parameters, toggling Boolean parameters and resampling categorical parameters;

generating the plurality of parameter sets based on the random changes;

generating a plurality of cluster solutions from each of the plurality of parameter sets, the plurality of cluster solutions generated for each parameter set comprising an optimized cluster solution having a highest total score of the plurality of cluster solutions generated for the parameter set;

obtaining a respective fitness score for each parameter set of the plurality of parameter sets, the respective fitness score of the parameter set being based on total scores of the plurality of cluster solutions generated for the parameter set;

selecting a hyper-optimized parameter set from the plurality of parameter sets based on a comparison of the respective fitness scores of the plurality of parameter sets;

selecting a hyper-optimized cluster solution from the plurality of cluster solutions generated from the hyper-optimized parameter set, the selection of the hyper-optimized cluster solution being based on comparisons of the total scores of the plurality of cluster solutions generated for the hyper-optimized parameter set; and

performing an actionable item that includes transmission of a communication by an external system to a group of users based on the hyper-optimized cluster solution.

9. The method of claim 8, wherein obtaining the fitness score further comprises:

selecting an upper quartile of total scores of the plurality of cluster solutions; and

calculating an average of the upper quartile total scores.

10. The method of claim 8, further comprising:

calculating respective total scores for the plurality of cluster solutions wherein each respective total score is a sum of a business value score, a solution diversity score and an average centroid distance score.

11. The method of claim 8, wherein a dataset associated with the plurality of parameter sets represents stores and the actionable item further comprises:

selecting a subset of the stores for upgrades.

12. A non-transitory computer readable medium including machine readable instructions that are executable by at least one processor to:

generate an initial parameter set of a plurality of parameter sets via random selection of parameter values, the initial parameter set including different types of parameters;

cause random changes in values of the different types of parameters in the initial parameter set, wherein the random changes include changes of a predetermined percentage caused in the values of numerical parameters, toggling Boolean parameters and resampling categorical parameters;

generate the plurality of parameter sets based on the random changes;

generate a plurality of cluster solutions from each of the plurality of parameter sets, the plurality of cluster solutions generated for each parameter set comprising an optimized cluster solution having a highest total score of the plurality of cluster solutions generated for the parameter set;

obtain a respective fitness score for each parameter set of the plurality of parameter sets, the respective fitness score of the parameter set being based on total scores of the plurality of cluster solutions generated for the parameter set;

select a hyper-optimized parameter set from the plurality of parameter sets based on a comparison of the respective fitness scores of the plurality of parameter sets;

select a hyper-optimized cluster solution from the plurality of cluster solutions generated from the hyper-optimized parameter set, the selection of the hyper-optimized cluster solution being based on comparisons of the total scores of the plurality of cluster solutions generated for the hyper-optimized parameter set; and

perform an actionable item that includes transmission of a communication by an external system to a group of users based on the hyper-optimized cluster solution.