Abstract: A data model is traversed to determine concept characteristics associated with concepts that may be associated with entities. Associated documents may be evaluated to identify document characteristics associated with the entities. Entity models may be trained based on the concept characteristics and the document characteristics with each entity model being associated with a confidence value. Results for one or more queries based on the documents and the entity models may be provided. The results may reference the documents that may be associated with the entities. Some entity models may produce results that have a confidence value below a threshold value. Accordingly, the entity models that provide low confidence results may be re-trained.

Abstract: Embodiments are directed towards managing data. Attributes of model fields of a plurality of model objects may be analyzed. If the analysis of the attributes discovers primary key fields in the model objects, the characteristics of the primary keys may be compared with the characteristics of other model objects. If affirmative results of the comparison indicate that one or more foreign key fields may be in the other model objects, one or more relationships that associate the primary key fields with the foreign key fields may be provided. And, a system model may be provided based on the relationships and the model objects that include the primary key fields and the model objects that include the foreign key fields.

Abstract: Embodiments are directed towards managing data using modeling platform. A principal data set and match data set may be provided. The principal data set may include principal objects and the match data set may include other model objects. Blocking may associate each of the plurality of principal objects with potential match objects. Principal objects from the plurality of principal objects may be selected based on heuristics, their associated potential match objects, or the like. Potential match objects may be labeled as a true match or a non-match with respect to their associated principal object. A ranker may be trained to identify top-ranked matches based on the labeled potential match objects. The trained ranker may be employed to indicate matches where each matching other model object is a potential match object that is ranked higher than other potential match objects that are associated with its matched principal object.

Abstract: Embodiments are directed towards managing data. An attributes engine may be employed to perform various actions, including: analyzing characteristics of model object features of a plurality of model objects; classifying the model object features based on the characteristics, such that the characteristics include a data type and values of the model object features; and associating similarity tasks with the model object features based on their classification. A similarity engine may then be employed to perform further actions, including: providing a similarity model that includes the similarity tasks; employing the similarity model to provide candidate similarity scores based on exemplar model objects labeled as being similar; modifying the similarity model based on the exemplar model objects and the candidate similarity scores; employing the modified similarity model to provide similarity scores for model objects based on the one or more similarity tasks.

Abstract: Embodiments are directed towards managing data using modeling platform. A principal data set and match data set may be provided. The principal data set may include principal objects and the match data set may include other model objects. Blocking may associate each of the plurality of principal objects with potential match objects. Principal objects from the plurality of principal objects may be selected based on heuristics, their associated potential match objects, or the like. Potential match objects may be labeled as a true match or a non-match with respect to their associated principal object. A ranker may be trained to identify top-ranked matches based on the labeled potential match objects. The trained ranker may be employed to indicate matches where each matching other model object is a potential match object that is ranked higher than other potential match objects that are associated with its matched principal object.

Abstract: Embodiments are directed towards dynamic semantic models having multiple indices. Source data may be provided to a network computer from at least one separate data source. A raw data graph may be generated from the source data such that the structure of the raw data graph may be based on the structure of the source data. Elements of the raw data graph may be mapped to a concept graph. Concept instances may be generated based on the concept graph, the raw data graph, and the source data. Model-identifiers (MIDs) that correspond to the concept instances may be generated to include at least a path in the concept graph The MID values may be indexed into a plurality of indices based on a content-type of the data associated with the MIDs. In response to a query, a result set may be generated that includes result MIDs.

Abstract: Embodiments are directed towards dynamic semantic models having multiple indices. Source data may be provided to a network computer from at least one separate data source. A raw data graph may be generated from the source data such that the structure of the raw data graph may be based on the structure of the source data. Elements of the raw data graph may be mapped to a concept graph. Concept instances may be generated based on the concept graph, the raw data graph, and the source data. Model-identifiers (MIDs) that correspond to the concept instances may be generated to include at least a path in the concept graph The MID values may be indexed into a plurality of indices based on a content-type of the data associated with the MIDs. In response to a query, a result set may be generated that includes result MIDs.