Cross-domain clusterability evaluation for cross-guided data clustering based on alignment between data domains

A computer program product, comprising a computer readable hardware memory having a computer readable program code embodied therein, said computer readable program code containing instructions that perform a method, said method comprising:

• creating, by a processor of a computer system, a cross-domain similarity graph over source centroids of a source domain and the target centroids of a target domain, using projected weights of pivot words in the source domain and the target domain, said cross-domain similarity graph comprising a first set of vertices corresponding to the source centroids and a second set of vertices corresponding to the target centroids, wherein a weight is assigned to each edge between one of the vertices in the first set vertices corresponding to the source centroids and one of the vertices in the second set vertices corresponding to the target centroids;
• said processor finding an alignment between the source centroids and the target centroids over the cross-domain similarity graph, said finding the alignment comprising determining a subset of the edges having a maximum sum of the weights as compared with a sum of the weights for all other subsets of the edges. subject to each vertex connected by the edges in the subset being spanned by at most one edge in the subset;
• said processor calculating target clusterability as an average of a respective clusterability of at least one target data item comprised by the target domain;
• said processor calculating target-side matchability as an average of a respective matchability of each target centroid of the target domain to source centroids of the source domain, wherein the source domain comprises at least one source data item;
• said processor calculating source-side matchability as an average of a respective matchability of each source centroid of said source centroids to the target centroids;
• said processor calculating source-target pair matchability as an average of the target-side matchability and the source-side matchability;
• said processor calculating cross-domain clusterability between the target domain and the source domain as a linear combination of the calculated target clusterability and the calculated source-target pair matchability, said cross-domain clusterability providing an improved clustering over conventional data clustering due to the alignment between the source centroids and the target centroids;
• and
• said processor transferring the calculated cross-domain clusterability to a first device selected from the group consisting of an output device of a computer system, a storage device of the computer system, a remote computersystem coupled to the computer system, and a combination thereof.

A computer system comprising a processor and a computer readable memory unit coupled to the processor, said computer readable memory unit containing instructions that when run by the processor implement a method, said method comprising:

• said processor creating a cross-domain similarity graph over source centroids of a source domain and the target centroids of a target domain, using projected weights of pivot words in the source domain and the target domain, said cross-domain similarity graph comprising a first set of vertices corresponding to the source centroids and a second set of vertices corresponding to the target centroids, wherein a weight is assigned to each edge between one of the vertices in the first set vertices corresponding to the source centroids and one of the vertices in the second set vertices corresponding to the target centroids;
• said processor finding an alignment between the source centroids and the target centroids over the cross-domain similarity graph, said finding the alignment comprising determining a subset of the edges having a maximum sum of the weights as compared with a sum of the weights for all other subsets of the edges, subject to each vertex connected by the edges in the subset being spanned by at most one edge in the subset;
• said processor calculating target clusterability as an average of a respective clusterability of at least one target data item comprised by the target domain;
• said processor calculating target-side matchability as an average of a respective matchability of each target centroid of the target domain to source centroids of the source domain, wherein the source domain comprises at least one source data item;
• said processor calculating source-side matchabillity as an average of a respective matchability of each source centroid of said source centroids to the target centroids;
• said processor calculating source-target pair matchability as an average of the target-side matchability and the source-side matchability;
• said processor calculating cross-domain clusterability between the target domain and the source domain as a linear combination of the calculated target clusterability and the calculated source-target pair matchability, said cross-domain clusterability providing an improved clustering over conventional data clustering due to the alignment between the source centroids and the target centroids;
• and
• said processor ansferring the calculated cross-domain clusterability to a first device selected from the group consisting of an output device of a computer system, a storage device of the computer system, a remote computer system coupled to the computer system, and a combination thereof.