# US 10,769,193 B2 · Predictive model clustering > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,769,193 B2 - Document ID: US10769193B2 - Status: Granted - Kind: B2 - Application number: 15/628,160 - Filed: 2017-06-20 - Published: 2018-12-20 - Granted: 2020-09-08 - Earliest claimed: 2017-06-20 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Eric Bouillet; Bei Chen; Randall L. Cogill; Thanh L. Hoang; Marco Laumanns; Karol W. Lynch; Rahul Nair; Pascal Pompey; John Sheehan - Web page: https://patentgrep.com/patent/US10769193B2 ## CPC - G06F 16/355: Creation or modification of classes or clusters - G06F 16/285: {Clustering or classification} - G06F 18/23: Clustering techniques - G06F 16/35: Clustering; Classification - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Performing data clustering in a model property vector space. Input data is received comprising a plurality of data instances in a data vector space. A model property vector specification is defined for a model vector. Information is identified from the input data, and a model property vector is created in the model property vector space for each of the plurality of data instances. A target number of clusters is identified and used to perform a data clustering procedure. An output is generated comprising a plurality of data segments and one or more clustering rules. For each data cluster, a predictive model is constructed for each data segment of the plurality of data segments. ## Claims ### Claim 1 A computer-implemented method for performing data clustering in a model property vector space, the method comprising: receiving input data from a database, the input data comprising a plurality of data instances in a data vector space, said plurality of data instances representing temporal features and geographical features information; defining a model property vector specification for creating a model vector in a model property vector space, the model property vector specification including one or more categories of information to be predicted for a predictive analytic problem; using the input data and the model property vector specification to identify information from the input data, the identified information comprising a temporal feature attribute, a geographic feature attribute, and a categorical attribute used to specify an item to be predicted for a predictive model; creating a model property vector in the model property vector space from the identified information for each data instance of the plurality of data instances, said model property vector space being separate from the data vector space and having data instances defining a sub-space of the data vector space; identifying a target number of clusters comprising a plurality of data clusters; using the target number of clusters and the data instances in the model property vector space to perform a data clustering procedure to cluster data in the model property vector space; generating an output from the data clustering procedure comprising a plurality of data segments and one or more clustering rules for the plurality of data clusters; and for each data cluster of the plurality of data clusters, constructing a predictive model for each data segment of the plurality of data segments. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, further comprising training the predictive model after the data clustering procedure is performed. ### Claim 3 (depends on 1) The computer-implemented method of claim 1 wherein the categorical attribute differentiates among a plurality of types or categories of prediction. ### Claim 4 (depends on 1) The computer-implemented method of claim 1, wherein the model property vector specification further indicates how categories of information to be predicted for is organized within the model property vector. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, wherein the temporal attribute information comprises a temporal prediction problem attribute, said prediction problem attribute comprising a prediction horizon. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein the geographical attribute information comprises a geographical attribute used to predict a geographic location. ### Claim 7 A computer program product for performing data clustering in a model property vector space, the computer program product comprising a computer-readable storage medium having a computer-readable program stored therein, wherein the computer-readable program, when executed on a processor, causes the processor to: receive input data from a database, the input data comprising a plurality of data instances in a data vector space, said plurality of data instances representing temporal features and geographical features information; define a model property vector specification for creating a model vector in a model property vector space, the model property vector specification including one or more categories of information to be predicted for a predictive analytic problem; use the input data and the model property vector specification to identify information from the input data, the identified information comprising a temporal feature attribute, a geographic feature attribute, and a categorical attribute information used to specify an item to be predicted for a predictive model; create a model property vector in the model property vector space from the identified information for each data instance of the plurality of data instances, said model property vector space being separate from the data vector space and having data instances defining a sub-space of the data vector space; identify a target number of clusters comprising a plurality of data clusters; use the target number of clusters and the data instances in the model property vector space to perform a data clustering procedure to cluster data in the model property vector space; generate an output from the data clustering procedure comprising a plurality of data segments and one or more clustering rules for the plurality of data clusters; and for each data cluster of the plurality of data clusters, construct a predictive model for each data segment of the plurality of data segments. ### Claim 8 (depends on 7) The computer program product of claim 7, further configured for training the predictive model after the data clustering procedure is performed. ### Claim 9 (depends on 8) The computer program product of claim 8 wherein the categorical attribute differentiates among a plurality of types or categories of prediction. ### Claim 10 (depends on 7) The computer program product of claim 7, wherein the model property vector specification further indicates how categories of information to be predicted for is organized within the model property vector. ### Claim 11 (depends on 7) The computer program product of claim 7, wherein the temporal attribute information comprises a temporal prediction problem attribute, said prediction problem attribute comprising a prediction horizon. ### Claim 12 (depends on 7) The computer program product of claim 7, wherein the geographical attribute information comprises a geographical attribute used to predict a geographic location. ### Claim 13 An apparatus for performing data clustering in a model property vector space, the apparatus comprising a processor and a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, provide: a model property vector creation module; a data clustering module operatively coupled to the model property vector creation module; and a predictive models training module operatively coupled to the data clustering module; wherein the model property vector creation module is configured to receive input data from a database, the input data comprising a plurality of data instances in a data vector space, said plurality of data instances representing temporal features and geographical features information; define a model property vector specification for creating a model vector in a model property vector space, the model property vector specification including one or more categories of information to be predicted for a predictive-analytic problem; use the input data and the model property vector specification to identify information from the input data, the identified information comprising a temporal feature attribute, a geographic feature attribute, and a categorical attribute information used to specify an item to be predicted for a predictive model; create a model property vector in the model property vector space from the identified information for each data instance of the plurality of data instances, said model property vector space being separate from the data vector space and having data instances defining a sub-space of the data vector space; and identify a target number of clusters comprising a plurality of data clusters; wherein the data clustering module is configured to use the target number of clusters and the data instances in the model property vector space to perform a data clustering procedure to cluster data in the model property vector space; and to generate an output from the data clustering procedure comprising a plurality of data segments and one or more clustering rules for the plurality of data clusters; and wherein the predictive models training module is configured, for each data cluster of the plurality of data clusters, to construct a predictive model for each data segment of the plurality of data segments. ### Claim 14 (depends on 13) The apparatus of claim 13, further configured for training the predictive model after the data clustering procedure is performed. ### Claim 15 (depends on 13) The apparatus of claim 13 wherein the categorical attribute differentiates among a plurality of types or categories of prediction. ### Claim 16 (depends on 13) The apparatus of claim 13, wherein the model property vector specification further indicates how categories of information to be predicted for is organized within the model property vector. ### Claim 17 (depends on 13) The apparatus of claim 13, wherein the temporal attribute information comprises a temporal prediction problem attribute, said prediction problem attribute comprising a prediction horizon.