# US 2021/0240727 A1 · Automatic feature engineering during online scoring phase > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2021/0240727 A1 - Document ID: US20210240727A1 - Status: Published application - Kind: A1 - Application number: 16/777,068 - Filed: 2020-01-30 - Published: 2021-08-05 - Earliest claimed: 2020-01-30 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Thanh Lam HOANG; Hong MIN - Web page: https://patentgrep.com/patent/US20210240727A1 ## CPC - G06F 16/26: Visual data mining; Browsing structured data - G06F 16/906: Clustering; Classification - G06F 16/2272: {Management thereof} - G06F 16/248: Presentation of query results - G06F 16/287: {Visualization; Browsing} - G06N 20/00: Machine learning - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Embodiments for intelligent automated feature engineering for relational data in a computing environment by a processor. Indices may be automatically selected and built from one or more columns of one or more tables in a relational database using one or more automated feature engineering models that include a set of queries. One or more features may be determined using a set of queries of an automated feature engineering models to execute for a scoring operation. ## Claims ### Claim 1 A method, by a processor, for intelligent automated feature engineering for relational data in a computing environment, comprising: automatically selecting and building indices from one or more columns of one or more tables in a relational database using one or more automated feature engineering models that include a set of queries; and determining one or more features using the set of queries of an automated feature engineering models to execute for a scoring operation. ### Claim 2 (depends on 1) The method of claim 1, further including selecting the one or more columns for selecting and building the indices according to the set of queries in the one or more automated feature engineering models. ### Claim 3 (depends on 1) The method of claim 1, further including selecting the one or more columns for selecting and building the indices according one or more constraints of a size of an index in the relational database. ### Claim 4 (depends on 1) The method of claim 1, further including maximizing a scoring query throughput based on one or more constraints in the relational database for selecting the one or more columns for selecting and building the indices. ### Claim 5 (depends on 1) The method of claim 1, further including automatically re-indexing data in the relational database to adjust a size of one or more sets of column indices according to one or more constraints. ### Claim 6 (depends on 1) The method of claim 1, further including automatically creating one or more queries to identify one or more key values matching records in the one or more tables. ### Claim 7 (depends on 1) The method of claim 1, further including: automatically creating one or more queries that performs a look-up for joined keys in the indices; and filtering contextual data using the joined keys for performing the scoring operation for joining with a main table, wherein the one or more tables include the main table and one or more contextual tables. ### Claim 8 A system for implementing intelligent automated feature engineering for relational data in a computing environment, comprising: one or more computers with executable instructions that when executed cause the system to: automatically select and build indices from one or more columns of one or more tables in a relational database using one or more automated feature engineering models that include a set of queries; and determine one or more features using the set of queries of an automated feature engineering models to execute for a scoring operation. ### Claim 9 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to select the one or more columns for selecting and building the indices according to the set of queries in the one or more automated feature engineering models. ### Claim 10 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to select the one or more columns for selecting and building the indices according one or more constraints of a size of an index in the relational database. ### Claim 11 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to maximize a scoring query throughput based on one or more constraints in the relational database for selecting the one or more columns for selecting and building the indices. ### Claim 12 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to automatically re-index data in the relational database to adjust a size of one or more sets of column indices according to one or more constraints. ### Claim 13 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to automatically create one or more queries to identify one or more key values matching records in the one or more tables. ### Claim 14 (depends on 8) The system of claim 8, wherein the executable instructions that when executed cause the system to: automatically create one or more queries that performs a look-up for joined keys in the indices; and filter contextual data using the joined keys for performing the scoring operation for joining with a main table, wherein the one or more tables include the main table and one or more contextual tables. ### Claim 15 A computer program product for, by a processor, implementing intelligent automated feature engineering for relational data in a computing environment, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising: an executable portion that automatically selects and builds indices from one or more columns of one or more tables in a relational database using one or more automated feature engineering models that include a set of queries; and an executable portion that determines one or more features using the set of queries of an automated feature engineering models to execute for a scoring operation. ### Claim 16 (depends on 15) The computer program product of claim 15, further including an executable portion that selects the one or more columns for selecting and building the indices according to the set of queries in the one or more automated feature engineering models. ### Claim 17 (depends on 15) The computer program product of claim 15, further including an executable portion that selects the one or more columns for selecting and building the indices according one or more constraints of a size of an index in the relational database. ### Claim 18 (depends on 15) The computer program product of claim 15, further including an executable portion that maximizes a scoring query throughput based on one or more constraints in the relational database for selecting the one or more columns for selecting and building the indices. ### Claim 19 (depends on 15) The computer program product of claim 15, further including an executable portion that: automatically re-indexes data in the relational database to adjust a size of one or more sets of column indices according to one or more constraints; or automatically creates one or more queries to identify one or more key values matching records in the one or more tables. ### Claim 20 (depends on 15) The computer program product of claim 15, further including an executable portion that: automatically creates one or more queries that performs a look-up for joined keys in the indices; and filters contextual data using the joined keys for performing the scoring operation for joining with a main table, wherein the one or more tables include the main table and one or more contextual tables.