Intelligent clustering systems and methods useful for domain protection

An intelligent clustering system, comprising:

• a processor;
• a non-transitory computer-readable medium;
• and
• instructions stored on the non-transitory computer-readable medium, wherein the instructions when translated by the processor implement a clustering engine and a prediction engine, wherein the clustering engine is operable to perform:
• analyzing a data set based on automatically-identified features, user-indicated features, or a combination thereof, the data set containing different domains associated with heterogeneous data types, wherein the analyzing comprises generating a tree data model for each of the different domains, the tree data model having nodes representing features and relationship paths representing non-linear relationships between the features;
• generating, based on the analyzing, domain clusters, wherein the generating comprises identifying common features between the different domains that share a high degree of feature overlap, wherein the generating comprises calculating a tree distance between the tree data models of the different domains based on weights associated with the nodes;
• and
• tuning the domain clusters, wherein the tuning comprises performing an iterative self-tuning process until a condition is met, wherein the iterative self-tuning process comprises:
• providing a clustering result to the prediction engine, wherein the prediction engine is operable to recognize domain elements found in new data and generate a cluster assignment or a cluster creation for each of the domain elements;
• updating the domain clusters with the domain elements based on the cluster assignment or the cluster creation;
• and
• providing an optimized clustering result to the prediction engine, wherein the prediction engine is configured for adjusting prediction parameters based on feedback received from the clustering engine through the optimized clustering result, and wherein the domain clusters are tuned in the stream-processing mode without waiting for completion of data collection of the data set.

A computer program product comprising a non-transitory computer-readable medium storing instructions, wherein the instructions when translated by a processor implement a clustering engine and a prediction engine, wherein the clustering engine is operable to perform:

• analyzing a data set based on automatically-identified features, user-indicated features, or a combination thereof, the data set containing different domains associated with heterogeneous data types, wherein the analyzing comprises generating a tree data model for each of the different domains, the tree data model having nodes representing features and relationship paths representing non-linear relationships between the features;
• generating, based on the analyzing, domain clusters, wherein the generating comprises identifying common features between the different domains that share a high degree of feature overlap, wherein the generating comprises calculating a tree distance between the tree data models of the different domains based on weights associated with the nodes;
• and
• tuning the domain clusters, wherein the tuning comprises performing an iterative self-tuning process until a condition is met, wherein the iterative self-tuning process comprises:
• providing a clustering result to the prediction engine, wherein the prediction engine is operable to recognize domain elements found in new data and generate a cluster assignment or a cluster creation for each of the domain elements;
• updating the domain clusters with the domain elements based on the cluster assignment or the cluster creation;
• and
• providing an optimized clustering result to the prediction engine, wherein the prediction engine is configured for adjusting prediction parameters based on feedback received from the clustering engine through the optimized clustering result, and wherein the domain clusters are tuned in the stream-processing mode without waiting for completion of data collection of the data set.