Abstract: Techniques and solutions are described for analyzing results of a machine learning model. A result is obtained for a data set that includes a first plurality of features. A plurality of feature groups are defined. At least one feature group contains a second plurality of features of the first plurality of features. The second plurality of features is less than all of the first plurality of features. Feature groups can be defined based on determining dependencies between features of the first plurality of features, including using contextual contribution values. Group contextual contribution values can be determined for feature groups by aggregating contextual contribution values of the constituent features of the feature groups.

Abstract: Techniques and solutions are described for analyzing results of a machine learning model. A result is obtained for a data set that includes a first plurality of features. A plurality of feature groups are defined. At least one feature group contains a second plurality of features of the first plurality of features. The second plurality of features is less than all of the first plurality of features. Feature groups can be defined based on determining dependencies between features of the first plurality of features, including using contextual contribution values. Group contextual contribution values can be determined for feature groups by aggregating contextual contribution values of the constituent features of the feature groups.

Abstract: Techniques and solutions are described for analyzing results of a machine learning model. A result is obtained for a data set that includes a first plurality of features. A plurality of feature groups are defined. At least one feature group contains a second plurality of features of the first plurality of features. The second plurality of features is less than all of the first plurality of features. Feature groups can be defined based on determining dependencies between features of the first plurality of features, including using contextual contribution values. Group contextual contribution values can be determined for feature groups by aggregating contextual contribution values of the constituent features of the feature groups.

Abstract: A method, a system, and a computer program product for determining uncertainties associated with a predictive modeling environment executed by computing systems. A dataset that includes a plurality of variables associated with one or more values is received for training a predictive model. The predictive model is trained using the received dataset and applied to one or more variables in the received dataset to generate a prediction. One or more uncertainty intervals corresponding to one or more contributions of one or more missing values corresponding to one or more variables in the plurality of variables are generated. One or more uncertainty intervals corresponding to one or more contributions of one or more rare values corresponding to one or more variables in the plurality of variables are generated. An alert indicative of the prediction is generated based on the one or more generated uncertainty intervals.

Abstract: Techniques and solutions are described for analyzing results of a machine learning model. A result is obtained for a data set that includes a first plurality of features. A plurality of feature groups are defined. At least one feature group contains a second plurality of features of the first plurality of features. The second plurality of features is less than all of the first plurality of features. Feature groups can be defined based on determining dependencies between features of the first plurality of features, including using contextual contribution values. Group contextual contribution values can be determined for feature groups by aggregating contextual contribution values of the constituent features of the feature groups.

Abstract: A method of improving usability and transparency of machine-learning aspects of applications providing various types of services is disclosed. Based on a request submitted through an administrative user interface, a data readiness check is performed on underlying data associated with the application. Based on a successful completion of the data readiness check, a configuration file is retrieved from an application server. The configuration file specifies a plurality of keys for generating a machine-learned model for the application. The machine-learned model is trained based on the plurality of keys specified in the configuration file. The machine-learned model is selected from a plurality of machine-learned models based on dry runs of the each of the plurality of models. The machine-learned model is activated with respect to the application. Scores are identified from the underlying data items based on the selected machine-learned model.

Abstract: A computer-implemented method for generating a plurality of data query suggestions is described. The method includes receiving a textual input in a user interface of a software application implementing a plurality of business processes and determining a query context for the textual input. The method also includes computing a plurality of ranked numerical scores based on the query context in which the ranked numerical scores are computed using information obtained from a plurality of usage metrics associated with the query context and determining, using the plurality of ranked numerical scores, a plurality of candidate data combinations. The method also includes ordering the plurality of candidate data combinations according to the plurality of ranked numerical scores and generating at least one data suggestion using the ordered plurality of candidate data combinations and providing the at least one data suggestion in the user interface.

Abstract: Embodiments visualize large data volumes utilizing initial sampling to reduce size of a dataset. This sampling may be random in nature. The sampled dataset may be refined (wrangled) by binning, grouping, cleansing, and/or other techniques to produce a wrangled sample dataset. A user defines useful end visualization(s) by inputting expected dimension/measures. From these visualizations of sampled data, minimal grouping sets are deduced for application to the full dataset. The user publishes/schedules the wrangled operation and grouping sets definition. Based on this, a wrangled dataset and grouping sets are produced in the big data layer. When the user accesses the visualization(s), minimal grouping sets are retrieved in the in-memory engine of the client and processed by an in-memory database engine according to the common processing plan. This produces result sets and a final set of visualizations of the full dataset, in which the user can recognize valuable data trends and/or relationships.

Abstract: A computer-implemented method for generating a plurality of data suggestions is described. The method includes receiving a textual input in a user interface of a software application implementing a plurality of business processes, determining a query context associated with the textual input, a user, and data that the user is viewing in the software application, and computing a plurality of statistical metrics for the query context, the statistical metrics being computed using information obtained from datasets associated with the query context. The method also includes determining a plurality of candidate data combinations, the data combinations including a plurality of dimensions, measures, and filters compatible with the query context, ranking the plurality of candidate data combinations according to one of the plurality of statistical metrics, and generating at least one data suggestion using the plurality of candidate data combinations and providing the at least one data suggestion in the user interface.

Abstract: A system comprises reception of a dataset comprising at least one dataset attribute and one or more instances of each dataset attribute, determination of a similarity value between instances of each pair of one of the at least one dataset attribute and one of a plurality of stored attributes, determination of an intersection between instances of each pair of one of the at least one dataset attribute and one of the stored attributes having a similarity greater than a threshold, and generation of metadata associating a dataset attribute with a semantic entity with which a stored attribute having a similarity to the dataset attribute greater than the threshold is associated.

Abstract: A system comprises reception of a dataset comprising at least one dataset attribute and one or more instances of each dataset attribute, determination of a similarity value between instances of each pair of one of the at least one dataset attribute and one of a plurality of stored attributes, determination of an intersection between instances of each pair of one of the at least one dataset attribute and one of the stored attributes having a similarity greater than a threshold, and generation of metadata associating a dataset attribute with a semantic entity with which a stored attribute having a similarity to the dataset attribute greater than the threshold is associated.

Abstract: The present disclosure describes methods, systems, and computer program products for dynamic generation and modification of input schemas. One computer-implemented method includes receiving event data from a data source with an event collector, the data source registered with the event collector to transmit event data to the event collector, transmitting the received event data for analysis, saving the transmitted event data, determining whether to create or update a business intelligence view, determining whether to create or update an input schema based on the saved event data, and providing, by a computer, an input schema describing how data is represented in the database to a query generation engine.

Abstract: A computer-implemented method for generating a plurality of data query suggestions is described. The method includes receiving a textual input in a user interface of a software application implementing a plurality of business processes and determining a query context for the textual input. The method also includes computing a plurality of ranked numerical scores based on the query context in which the ranked numerical scores are computed using information obtained from a plurality of usage metrics associated with the query context and determining, using the plurality of ranked numerical scores, a plurality of candidate data combinations. The method also includes ordering the plurality of candidate data combinations according to the plurality of ranked numerical scores and generating at least one data suggestion using the ordered plurality of candidate data combinations and providing the at least one data suggestion in the user interface.

Abstract: A computer-implemented method for generating a plurality of data suggestions is described. The method includes receiving a textual input in a user interface of a software application implementing a plurality of business processes, determining a query context associated with the textual input, a user, and data that the user is viewing in the software application, and computing a plurality of statistical metrics for the query context, the statistical metrics being computed using information obtained from datasets associated with the query context. The method also includes determining a plurality of candidate data combinations, the data combinations including a plurality of dimensions, measures, and filters compatible with the query context, ranking the plurality of candidate data combinations according to one of the plurality of statistical metrics, and generating at least one data suggestion using the plurality of candidate data combinations and providing the at least one data suggestion in the user interface.

Abstract: The present disclosure describes methods, systems, and computer program products for enabling extreme visualization for large data sets stored in a database. One computer-implemented method includes receiving a request for an input schema to feed to a data specification language (DaSL) and a metadata outline of typed objects exposed by a database view identified in the request, requesting extraction of the input schema and metadata outline from the database, creating an instance of a DaSL compiler based upon the input schema, generating, by a computer, a DaSL query corresponding to objects selected from the metadata outline, requesting compilation of the DaSL query, and receiving a calculation plan and topology cursors responsive to the compilation of the DaSL query.

Abstract: Embodiments visualize large data volumes utilizing initial sampling to reduce size of a dataset. This sampling may be random in nature. The sampled dataset may be refined (wrangled) by binning, grouping, cleansing, and/or other techniques to produce a wrangled sample dataset. A user defines useful end visualization(s) by inputting expected dimension/measures. From these visualizations of sampled data, minimal grouping sets are deduced for application to the full dataset. The user publishes/schedules the wrangled operation and grouping sets definition. Based on this, a wrangled dataset and grouping sets are produced in the big data layer. When the user accesses the visualization(s), minimal grouping sets are retrieved in the in-memory engine of the client and processed by an in-memory database engine according to the common processing plan. This produces result sets and a final set of visualizations of the full dataset, in which the user can recognize valuable data trends and/or relationships.

Abstract: The present disclosure describes methods, systems, and computer program products for dynamic generation and modification of input schemas. One computer-implemented method includes receiving event data from a data source with an event collector, the data source registered with the event collector to transmit event data to the event collector, transmitting the received event data for analysis, saving the transmitted event data, determining whether to create or update a business intelligence view, determining whether to create or update an input schema based on the saved event data, and providing, by a computer, an input schema describing how data is represented in the database to a query generation engine.

Abstract: The present disclosure describes methods, systems, and computer program products for enabling extreme visualization for large data sets stored in a database. One computer-implemented method includes receiving a request for an input schema to feed to a data specification language (DaSL) and a metadata outline of typed objects exposed by a database view identified in the request, requesting extraction of the input schema and metadata outline from the database, creating an instance of a DaSL compiler based upon the input schema, generating, by a computer, a DaSL query corresponding to objects selected from the metadata outline, requesting compilation of the DaSL query, and receiving a calculation plan and topology cursors responsive to the compilation of the DaSL query.

Abstract: A system includes reception of a logical schema associated with a data store, generation of an input schema based on the logical schema, reception of an expression having a type of the input schema and representing the topology of a data set, the data set being an instance of the input schema, compilation of the expression to a data specification based on the input schema, and generation of a plan to execute on the data store based on the data specification.

Abstract: A system includes reception of a logical schema associated with a data store, generation of an input schema based on the logical schema, reception of an expression having a type of the input schema and representing the topology of a data set, the data set being an instance of the input schema, compilation of the expression to a data specification based on the input schema, and generation of a plan to execute on the data store based on the data specification.