Improved entity resolution for master data with qualified relationship scores

Abstract

A first score associated with a match between entity records of a plurality of entities in master data of an MDM system is received. A set of entity records having a first score above a lower threshold score and below an upper threshold score is not confirmed as a match or a non-match and is identified as unresolved. A second score associated with the relationship between the entity records is generated. A total score for the pairing of the set of entity records is determined by combining the first match score with the second relationship score. The total score for each pairing of the set of entities is compared to an upper threshold, and if the total score exceeds the upper threshold, information for the pair of entity records in the set of entity records is combined into a single record and redundant entity records are deleted from the MDM system.

Description

[Technical field]

[0001] The present invention relates generally to the field of master data management, and more specifically to improved resolution of entity data in master data management systems using modified relationship scores. [Background technology]

[0002] Master Data Management (MDM) is the core process used to manage, centralize, organize, classify, locate, synchronize and enrich master data according to the business rules of an enterprise's sales, marketing and operational strategies. MDM uses a comprehensive approach to define and manage an organization's critical data. MDM provides a single, trusted view of data across the organization's constituents, interactions and interests.

[0003] Master data includes data and information associated with entities with which an organization functions, supports or interacts, such as customers, clients, employees, contractors, patients, healthcare providers, service providers (people), other organizations with unique identifying attributes, locations, products or objects.

[0004] Data quality and accuracy are important attributes of a master data set, and MDM systems include techniques to avoid false positives and false negatives when resolving two entity records as the same entity. Summary of the Invention

[0005] An embodiment of the present invention discloses a method, a computer program product, and a system. The embodiment includes a method for resolving entity records in a master data management (MDM) system. The method provides one or more processors for receiving first scores associated with matches between entity records of a plurality of entities of master data of the MDM system. The one or more processors identify first scores of the set of entity records that are above a lower threshold matching score and below an upper threshold matching score, respectively, to indicate unresolved matches of the set of entity records. The one or more processors generate second scores associated with relationships between the entity records based on an analysis of relationship data associated with the set of entity records. The one or more processors generate a total score for the set of entity records by combining the first scores associated with the matches and the second scores associated with the relationships. The one or more processors determine whether a total score for the set of entities exceeds an upper threshold, and in response to a total score for a pairing of entity records of the set of entity records exceeding an upper threshold, the one or more processors combine information of the pair of entity records of the set of entity records into a single record and remove redundant entity records from the MDM system. [Brief description of the drawings]

[0006] [Figure 1] FIG. 1 is a functional block diagram illustrating a distributed data processing environment according to one embodiment of the present invention.

[0007] [Figure 2A] 1 illustrates categorization of entity record probabilistic matching according to an embodiment of the present invention.

[0008] [Figure 2B] 1 illustrates an example of weighted scores for entity relationship types according to an embodiment of the present invention.

[0009] [Diagram 3] 2 is a flow chart illustrating the operational steps of a Record Linking program operating in the distributed data processing environment of FIG. 1 in accordance with an embodiment of the present invention.

[0010] [Figure 4] 4 illustrates a block diagram of components of a computing system including a computing device operatively configured to execute the record linking program of FIG. 3, according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Embodiments of the present invention recognize that master data management (MDM) systems often utilize a probabilistic matching engine (PME) to evaluate data records of entities and determine whether the data records are duplicates of the entity or whether the data records are for separate entities. Embodiments also recognize that the PME solution creates a matching score in pairing each master data record with other master data records to determine the matching score based on attributes of each entity record. Embodiments recognize that the PME's matching algorithms are carefully calibrated to minimize false positive and false negative matches of entity records, as false matches can lead to highly problematic behavior and decisions.

[0012] In embodiments of the present invention, the term "entity" is used to refer to an object, and an "entity record" refers to a collection of data and information associated with an object. An entity may be, for example, an individual, a family, a group of people, a business, an organization, a location area, a product, a service, or any object having attributes or characteristics that provide an identification or description by the collection of data and information about the object. Individual people entities may be, for example, customers, patients, contacts, employees, members, and providers. Embodiments also recognize that an entity record may contain limited, changed, or erroneous data and information about the attributes of an entity, creating conditions where duplicate or redundant entity records may exist in the master data.

[0013] An embodiment of the present invention recognizes that a PME solution applied to an MDM system entity record results in an entity record score. An entity record contains data and information associated with attributes of an entity, which may include, for example, name or label, address, phone number, gender, and other attributes and characteristics. PME entity record matching generates a probability score by comparing available entity attributes between entity record pairs and assigning a high score to attribute matches and a low or negative score to attribute non-matches. The total score is applied to a pair of sorting thresholds. These two thresholds define limits that automatically determine whether a pair of records are from the same entity (above an upper threshold score) or from separate entities (below a lower threshold score). Record comparisons for which the PME solution generates a matching score above or below the lower threshold or below the upper threshold require additional resolution to determine whether the records are from the same entity or not.

[0014] Users of master data repositories rely on the high quality and accuracy of the data and often utilize manual review techniques performed by data stewards to resolve distinctions in entity records. Embodiments recognize that applying a conservative PME solution algorithm to entity records often results in a set of unresolved entity records that require manual review to resolve. In some cases, to avoid the costs associated with manual resolution through clerical review, the unresolved records are ignored and the inaccurate data is accepted.

[0015] Embodiments of the present invention provide methods, computer program products and computer systems for resolving potential duplicate or redundant entity records in master data. In embodiments of the present invention, an evaluation of relationship record data, grouping record data and hierarchy record data between pairs of entity records in master data is performed. This evaluation generates a relationship score that is added to the matching score of the PME algorithm, resulting in a total score to more accurately determine whether the record pairs are of the same entity or of separate entities. In some embodiments, the relationship attributes compared between the entities include common grouping and hierarchical relationships, such as reporting structures or functional interactions within a business or organization. Embodiments of the present invention refer herein to entity record attributes of relationships as including grouping and hierarchy attributes between entities.

[0016] An embodiment of the present invention recognizes that an MDM system utilizing a PME algorithm solution includes only category and demographic data for entity matching and resolution. An embodiment of the present invention includes the addition of relationship data that is added to the MDM system's existing category and demographic matching data. In some embodiments, a second score is generated that reflects the relationship, grouping, and hierarchy information associated with the master data entity. The second "relationship" score is combined with the initial "matching score" to provide a total score for the entity pairing. The total score is then compared to upper and lower thresholds to determine whether the entity pair is a match, a mismatch, or part of a significantly reduced unresolved category.

[0017] The present invention will now be described in detail with reference to the figures. Figure 1 is a functional block diagram illustrating a distributed data processing environment, generally designated 100, in accordance with one embodiment of the present invention. Figure 1 is provided only as an illustration of one implementation and is not intended to suggest any limitations with respect to the environments in which different embodiments may be implemented. One of ordinary skill in the art may make numerous modifications to the depicted environment without departing from the scope of the present invention as set forth in the claims.

[0018] The distributed data processing environment 100 includes a master data management (MDM) system 110, a server 120, category and demographic (C&D) information 130, relationship information 140, linked entity records 160, and unresolved entity records 170, all interconnected via a network 150. The network 150 may be, for example, a local area network (LAN), the Internet, a wide area network (WAN) such as a virtual local area network (VLAN), or any combination that may include wired, wireless, or optical connections. In general, the network 150 may be any combination of supporting connections and protocols.

[0019] The MDM system 110 includes master data 115 and a probabilistic matching engine (PME) 117. The MDM system 110 receives, stores, processes, and updates the master data 115. The MDM system 110 may include a combination of computing devices, processing devices, storage devices, editing devices, and applications for maintaining the master data 115. The MDM system 110 performs matching operations via the PME 117, which compares each entity record with every other entity record in the master data 115 to determine whether the records are of the same entity or of separate entities. The matching operations performed by the MDM system 110 include generating a first score, referred to herein as a matching score, based on the probability that the attributes of the two entity records are of the same entity.

[0020] Master data 115 is shown as a component of MDM system 110. In some embodiments, master data 115 may be included in a searchable database or repository and may be an integrated part of MDM system 110. In other embodiments, master data 115 is a separate data repository (not shown) accessible to MDM system 110. Master data 115 includes information about attributes of entities in master data 115. In some embodiments, master data 115 includes attributes associated with each entity in master data 115, such as names or labels associated with that entity (e.g., customer, organization, etc.), location, subject matter, and other categorical and / or demographic data. PME 117 can utilize master data 115 to generate a first score (matching score) between pairs of entity records. In embodiments of the present application, master data 115 includes information and data from categorical and demographic sources, such as C&D information 130, as well as an additional set of relationship data, such as information included in relationship information 140.

[0021] C&D information 130 includes one or more sources of categorical and demographic entity data and information. In some embodiments, C&D information 130 includes data and information provided by the entity or the entity's support services. In other embodiments, C&D information 130 results from the entity's actions or transactions. An embodiment of the present invention considers data that can be described by a categorical description (age range, height, location, and descriptive information) to be categorical information. An embodiment considers demographic data to include names or labels, locations (of people or organizations), phone numbers, identification numbers, etc. C&D information 130 is utilized by PME 117 of MDM system 110 to generate a first score matching between entities in master data 115.

[0022] The relationship information 140 includes one or more sources of information regarding the relationship of an entity of the master data 115 to other entities, groups of entities, and hierarchical relationships with other entities. For example, the relationship information 140 may include information about a first entity (individual) having two siblings, an older brother and a younger sister, with the same first name, and a parent. In another example, the relationship information 140 includes information about a recreation organization of which the first user is a member, the entity in which the first user works, and the position of the first user within the entity. An embodiment of the present invention generates a second score utilizing the relationship information 140 for each pairing of entities of the master data 115 combined with the first score, which is the matching score performed by the PME 117, to generate an overall score for determining whether the pairing of entity records of the master data 115 is the same or distinct.

[0023] In various embodiments of the present invention, the categorical and demographic entity data and information included in C&D information 130 and the relationship data and information included in relationship information 140 are collected or provided by entities that have the option to "opt-in" or "opt-out." Embodiments of the present invention provide the entities with an explanation of how the collected or provided information will be used and allow the entities to permit or deny use of the collected or provided information for the described purposes.

[0024] Matching entity records 160 include all records originally classified as unresolved entity records after the first score of matching by PME 117 that are resolved as matching after obtaining a total score by combining the relationship score with the matching score. Matching entity records 160 are identified and sent to MDM system 110 for resolving duplicate entity records and combining previously separate entity records under a determined single entity. In some embodiments, data of entity record pairs are combined and verified inconsistent data and duplicate records are removed from the master data.

[0025] Unresolved entity records 170 include entity records whose total score after combining the matching scores with the relationship scores remains above a lower threshold score and below an upper threshold score. Unresolved entity records 170 remain unresolved and, in some embodiments, a data steward is assigned to complete resolution of a significantly reduced number of unresolved entity records.

[0026] The server 120 is shown as including a record linking program 300. The server 120 communicates with the MDM system 110 via the network 150 and forwards unresolved entity records output from the PME 117 to the record linking program 300. In some embodiments, the server 120 may be a web server, a blade server, a desktop computer, a laptop computer, a tablet computer, a netbook computer, or any other programmable electronic computing device capable of receiving, transmitting, and processing data within the distributed data processing environment 100 and communicating with the MDM system 110 via the network 150. In another embodiment, the server 120 represents a computing system utilizing clustered computers and components (e.g., database server computers, application server computers, etc.) that operate as a single pool of seamless resources when accessed within the distributed data processing environment 100. As shown and described in more detail in FIG. 4, the server 120 may include internal and external hardware components.

[0027] The record linking program 300 is shown running on the server 120 and communicatively connected to the MDM system 110 via the network 150. In some embodiments, the record linking program 300 may run within a computer server system (not shown) that includes the MDM system 110. The record linking program 300 receives initial matching scores (first scores) between pairings of unresolved entity records, where the matching scores are determined based on the categories and demographic data of the master data 115, excluding relationship information such as the relationship information 140. After identifying the set of unresolved entity records, the record linking program 300 determines the relationships between the pairings in which each of the unresolved entity records is paired with each of the other unresolved entity records. The record linking program 300 generates a relationship score (second score) based on the type of relationship, including the relationship between the paired entities and the relationships with other entity records. The record linking program 300 combines the relationship scores with the matching scores previously determined by the PME 117 of the MDM system 110 to generate a total score.

[0028] A lower threshold score and an upper threshold score are defined within the PME 117 and are used to indicate entity records as unmatched, unresolved, and matched. The record linking program 300 receives a matching score for each pairing of entity records that corresponds to the unresolved entity records that is above the lower threshold score but below the upper threshold score. After an initial load of the entities and their corresponding category and demographic data and information, the record linking program 300 generates a second score for each pairing of each unresolved entity record based on the relationship data added to the master data 115 of the MDM system 110.

[0029] In some embodiments, the second score is weighted based on the relationship data that more conclusively indicates that the paired entity records are of the same entity (higher positive score) or of separate entities (lower or negative score). The record linking program 300 combines the first score and the second score for each of the unresolved entity record pairings to generate a total score. The record linking program 300 compares the total score of each pairing of entity records to an upper score threshold and a lower score threshold. If the total score of the entity record pairing exceeds the upper score threshold, the record linking program 300 considers the record pair as a match and combines or merges the information of these two separate records into a single record, and removes redundant records and information from the master data 115. If the total score of the entity record pairing under consideration is below the lower threshold score, the record linking program 300 considers the pair or entity records as unmatched and separate records. If the total score of the pairing of entity records under consideration falls within or is between the lower and upper threshold scores, the record linking program 300 considers the pair or entity records under consideration to be unresolved, requiring further investigation, often manually performed by a data steward.

[0030] In some embodiments of the present invention, the record linking program 300 generates the second score based on relationships between the entities, groups of the entities, and hierarchical structures between the entities. For example, an individual may have a sibling, descendant, parent, or broader family relationship, or a combination thereof, may be a member of an organization with the other entity, may be in a common area with the other entity, may work in a common business organization with the other entity, and may hold a position that has a hierarchical structure related to the other entity. The presence or absence of one or more relationships of an unresolved entity record with another entity may increase, decrease, or have no effect on the second score of the pairing of each unresolved entity record based on a predetermined weight given to the particular state of a relationship.

[0031] In another exemplary embodiment, if two entity records indicate having a single sibling and each entity record contains a parent relationship common to both entity records, then the two entities are most likely to be separate and the relationship weights will be lower in value and potentially negative relative to the second score, resulting in a total score below the lower threshold score indicating the entity records do not match. If the two entities have common organizational memberships of spouses and a common name, then the two entity records are most likely to be the same entity and the scores for these relationships will be higher in value and positive indicating the two entity records match.

[0032] In some embodiments of the present invention, the record linking program 300 is a machine learning model to which the unresolved entity records are applied after initial matching scoring. In some embodiments, the machine learning model is trained by determining weighting coefficients corresponding to the relationship type, hierarchy condition, and common grouping attribute of each pair of unresolved entity records of master data that are manually resolved by a data steward, thus acting as a supervised learning of the machine learning model. After deployment of the record linking program 300 as a machine learning model, the record linking program 300 obtains results from the reduced but remaining unresolved entity records that are resolved by the data steward, thus continuously improving the machine learning model.

[0033] FIG. 2A illustrates categorization of entity record probabilistic matching according to an embodiment of the present invention. FIG. 2A includes unmatched records 210, unresolved records 215, matching records 220, lower threshold score 225, and upper threshold score 230. In an embodiment of the present invention, the entity records are classified into three conditions of entity record matching after the first score between the pairing of the master data entity records. The first score of the entity record pairs is performed by the PME 117 (FIG. 1). The categorization of the master data entity records is based on the first score of each entity record pairing with each of the other entity records, which may be referred to as the two entity records under review. The lower threshold score 225 and upper threshold score 230 are defined by the MDM system 110 and are coordinated with the scoring scheme and conservative objectives of the MDM system 110 to avoid false negatives and false positives that typically result in significant entity records being classified in the unresolved records 215.

[0034] A PME 117 operating on the master data management system 110 performs a comparison of entity records to determine a first score. The PME 117 compares an entity record to another entity record and scores the attributes of the entity records based on whether they match or nearly match, such that a match of attributes increases the first score and no match affects the first score or decreases the first score. The PME 117 performs the comparison iteratively for a combination of multiple entity records. The PME 117 compares the first score to a lower threshold score 225 and an upper threshold score 230 to determine a categorization of the entity record pair. If the PME 117 determines that the first score of the entity record is below the lower threshold score 225, it classifies the entity record as an unmatched record 210. The PME 117 classifies the entity record pair as a matching record 220 if the PME 117 determines that the first score exceeds the upper threshold score 230, and classifies the entity record pair as an unresolved record 215 if the first score is at or between the lower threshold score 225 and the upper threshold score 230.

[0035] FIG. 2B illustrates a table including exemplary weighted scores of entity relationship types according to an embodiment of the present invention. FIG. 2B includes sibling relationship type 240, parent relationship type 242, spouse relationship type 244, and employer relationship type 246 as respective columns of a relationship score table 270. The relationship score table 270 includes a row 250 including example relationship scores (second scores) of two entity records of unresolved master data under review in the format of "a, b." "a" is the weight of the relationship between the two entity records under review, and "b" is the weight of the absence of a relationship. FIG. 2B includes a row 260 representing example relationship scores of two entity records under review in the format of "x, y, z." "x" is the weight value of the relationship of the pair of entity records with another common entity, "y" is the weight value of the relationship of the pair of entity records with a different other entity, and "z" is the weight value indicating the absence of a relationship with the other entity.

[0036] Row 250 and sibling relationship type 240 indicate the weight of the relationship score for the sibling relationship between the pair of entity records under review. A large negative weight is therefore assigned (position "a") because it is unlikely that the two entity records are distinct. The weight indicating the absence of relationship "b" is maintained at a zero value when the pair of entity records under review are confirming sibling relationships.

[0037] Row 250 and parent relationship type 242 show example weights for relationship scores for parent-child relationships between pairs of entity records, and row 250 and spouse relationship type 244 show example weights for relationship scores for two entity records that have a spouse-spouse relationship. If the two entity records under review contain a parent or spouse relationship, then these entity records are most likely not the same entity and are assigned a large negative weight on the relationship score (second score) to influence the total score below the lower threshold score. Row 250 and employer relationship type 246 shows the relationship between the pair of entity records as having a common employer. The weights assigned to the example relationship scores are zero, meaning there is no weighted influence on the relationship score, since there is some likelihood that separate entities may work for the same employer.

[0038] Row 260 shows exemplary weights for the second score (relationship score) based on sibling, parent, spouse, and employer relationships. Row 260 and sibling relationship type 240 indicates a common sibling relationship of the pair of entity records with another entity (position "x") and receives a lower weight of 10 because the two records may be of the same entity or there may be a third sibling related to both. Row 260 and parent relationship type 242 indicates a common parent relationship of the pair of entity records with another entity (position "y") and receives a weight of 15 because the pair of records may be of the same entity or there may be siblings. Row 260 and spouse relationship type 244 indicates a common spouse relationship and receives a higher weight of 50 because it is most certain that the pair of entity records are the same entity. Row 260 and employer relationship type 246 shows exemplary weights for the second score based on the relationship of the entity record under review to the employer.

[0039] The example weights in Figures 2A and 2B are intended to reflect relationship considerations in determining whether paired entity records are likely to be the same entity and result in the records being a match, or whether these entity records are distinct. The magnitude of the weights is an example, and implementations of embodiments of the present invention include specific master data management considerations.

[0040] FIG. 3 is a flow chart illustrating the operational steps of a record linking program 300 operating in the distributed data processing environment of FIG. 1, in accordance with an embodiment of the present invention.

[0041] The record linking program 300 receives a first score associated with a match between each entity record in the master data (step 310). In some embodiments of the invention, entity records in the master data that have common or similar attributes are paired and scored (first score) by a probabilistic matching engine to determine whether the pair of entity records are from the same entity, from separate entities, or whether it is ambiguous (considered unresolved) as to whether the entity records are from the same entity or from different entities.

[0042] For example, pairs of master data entity records having the same surname attribute are processed by the probabilistic matching engine 117 of the MDM system 110 to yield a first score of a match between the pair of entity records.

[0043] The record linking program 300 identifies a set of unresolved entity records (step 320). In an embodiment of the invention, the record linking program 300 receives entity records classified as "unresolved" after a first score match to entity record pairs that were not considered record matches or separate records. In some embodiments, upper and lower threshold scores are established and the probabilistic matching engine applies the threshold scores to determine matching and non-matching entity record pairs. The upper and lower threshold scores are conservatively determined to avoid false positive and false negative matching of entity records. Conservative thresholds result in a significant amount of entity records being classified as "unresolved" after the initial matching score determination. The record linking program 300 determines the unresolved entity records as a set of additional matching resolutions.

[0044] For example, a number of data and information records of the master data 115 corresponding to a number of entities are processed by the PME 117 of the MDM system 110. The PME 117 generates a record matching probability score (first score) for pairs of entity records and classifies these records as "matching" based on the matching score of the pair of entity records exceeding the upper threshold score. The PME 117 can classify the appearance of entity records as "not matching" based on the matching score falling below the lower threshold score, and the pair of entity records having an initial matching score that is at or between the lower and upper threshold scores are classified as "unresolved". The record linking program 300 identifies the entity records having an initial matching score that is at or between the lower and upper threshold scores as a set of unresolved entity records.

[0045] The record linking program 300 generates a second score associated with the relationship of the unresolved entity record pairing (step 330). The master data management (MDM) system considers basic categorical and identifying information for the resolution of entity records, e.g., name, address / location, phone number, account, date, etc. In an embodiment of the invention, the MDM system includes qualified relationship information, which is used to generate the second score associated with the relationship between the entity records and between the pair of master data entity records and other entity records. An embodiment of the invention includes known qualified information associated with entity hierarchies and groups as indicated in the master data entity records in the relationship scores.

[0046] The record linking program 300 applies weightings to the matching scores of unresolved entity record pairs. In some embodiments, a positive weighting is applied to relationship, hierarchy, and grouping information that supports that the entity records may be from the same entity and may be a match. In some embodiments, the relationship information may strongly support a matching set of entity records and is therefore weighted significantly higher. In other embodiments, the relationship information indicates that the records are most likely from separate entities and is assigned a zero or negative weight value. In some embodiments, the record linking program 300 considers the relationship, hierarchy, and grouping information (collectively referred to herein as "relationship information") of the unresolved entity record pair under review along with one or more other entity records in determining the relationship score (second score).

[0047] For example, the record linking program 300 determines that each entity record of a pair of entity records is linked to a different employer and has a different spousal relationship. Because these records are distinct, the record linking program 300 applies a large negative weight (e.g., −100) to the pairing of records. Another pair of entity records has a sibling relationship with a common entity and has a common parent relationship. Because these records may be from the same entity, the record linking program 300 assigns these records a positive weight of 10. However, because there may be a third sibling to which two of the entity records are related, the assigned weight is somewhat less in value. In some embodiments, for faster determination and selection of relationships, the relationship information may be indexed or bucketed based on "entity-from-id", "entity-to-id" and "relationship type".

[0048] The record linking program 300 generates a total score by combining the first score and the second score (step 340). The record linking program 300 combines the first score generated by the initial probabilistic matching engine with the second score generated from the relationship information to generate a total score for the pairing of unresolved entity records under review. In some embodiments, the total score is the arithmetic sum of the matching score and the relationship score. In other embodiments, the relationship score may include a positive or negative factor multiplied by the first score matching result, such as a factor of -1.25 to reduce the first score by 25%.

[0049] The record linking program 300 determines whether the total score matching pair of entity records exceeds the upper threshold score (determination step 350). The record linking program 300 compares the total matching score of the pair of entity records to a previously established upper threshold score to determine whether the total score exceeds the upper threshold. In some embodiments of the invention, the relationship score raises the total matching score of the pair of entity records that exceeds the upper threshold score, while in other embodiments, the relationship score is significantly negative and the total matching score falls below the lower threshold score.

[0050] If the total score does not exceed the upper threshold score (step 350, "No" branch), the record linking program 300 proceeds to determine whether the total score is below the lower threshold score (decision step 370). In step 370, the record linking program 300 compares the total score of the matching pair of unresolved entity records against the lower threshold score to determine whether the total score is below the lower threshold score.

[0051] If the total score is below the lower threshold score (step 370, "Yes" branch), the record linking program 300 classifies the entity record pair as "non-matching" (step 380). When the record linking program 300 determines that the total score of matches for the unresolved entity record pair is below the lower threshold score, it classifies the entity record pair as "non-matching" and concludes that the records are from separate entities.

[0052] If the total score is greater than or equal to the lower threshold score, but the previous step 350 determined that the total score is less than or equal to the upper threshold score (step 370, "No" branch), the record linking program 300 classifies the entity record pair as remaining "unresolved" (step 390) and further investigation is required to resolve the remaining unresolved entity records from the total score. After comparing the total scores of the entity record pairings and determining the matching categories of the entity records, the record linking program 300 ends.

[0053] Returning to step 350, if the record linking program 300 determines that the total score exceeds the upper threshold score (step 350, "Yes" branch), then the record linking program 300 combines the information from each of the pair of entity records into a single entity record (step 360). The record linking program 300 combines the information from the entity records under review into a single record and removes redundant information and duplicate records from the master data.

[0054] For example, the record linking program 300 determines that the phone number and address of a more recent entity record have replaced the phone number and address of an older entity record. The record linking program 300 adds the new information about the entity's group membership and date of birth found in the more recent entity record to the updated entity record and discards the duplicate record attributes and entity records, leaving behind a single record for the entity with the combined and updated record attributes.

[0055] In some embodiments, additional resolution of unresolved entity records is performed manually by a data steward. In an embodiment of the invention, the inclusion of a second score based on known qualified relationship information of the entity for which the entity record is under review significantly reduces the amount of entity records remaining "unresolved," significantly reducing the time and cost of manual resolution by data stewards and avoiding loss of accuracy and quality of master data records by ignoring unresolved entity records.

[0056] FIG. 4 illustrates a block diagram of components of a computing system including a computing device 405 that includes or is configured to operatively connect to the components illustrated in FIG. 1 and is capable of operatively executing the record linking program 300 of FIG. 3 in accordance with one embodiment of the present invention.

[0057] In accordance with an exemplary embodiment of the invention, computing device 405 includes components and functionality similar to those of server 120 (FIG. 1). It should be understood that FIG. 4 is intended to provide an illustration of only one implementation and is not intended to suggest any limitation with respect to the environments in which different embodiments may be implemented. Many modifications to the illustrated environment may be made.

[0058] Computing device 405 includes a communications fabric 402 that provides communication between computer processor 404, memory 406, persistent storage 408, communications unit 410, and input / output (I / O) interface 412. Communications fabric 402 may be implemented using any architecture designed to pass data and / or control information between processors (e.g., microprocessors, communications processors, network processors, etc.), system memory, peripheral devices, and any other hardware components in the system. For example, communications fabric 402 may be implemented using one or more buses.

[0059] The memory 406, cache memory 416 and persistent storage 408 are computer-readable storage media. In this embodiment, the memory 406 includes random access memory (RAM) 414. In general, the memory 406 may include any suitable volatile or non-volatile computer-readable storage medium.

[0060] In one embodiment, the record link program 300 is stored in persistent storage 408 for execution by one or more of the respective computer processors 404 via one or more of the memories 406. In this embodiment, persistent storage 408 includes a magnetic hard disk drive. Alternatively, or in addition to a magnetic hard disk drive, persistent storage 408 may include a solid state hard drive, a semiconductor storage device, a read only memory (ROM), an erasable programmable read only memory (EPROM), a flash memory, or any other computer readable storage medium capable of storing program instructions or digital information.

[0061] The media used by persistent storage 408 may be removable. For example, a removable hard drive may be used for persistent storage 408. Other examples include optical and magnetic disks, thumb drives, and smart cards that are inserted into a drive for transfer to another computer-readable storage medium that is also part of persistent storage 408.

[0062] The communications unit 410, in these examples, provides for communication with other data processing systems or devices, including resources of the distributed data processing environment 100. In these examples, the communications unit 410 includes one or more network interface cards. The communications unit 410 may provide communications through the use of either or both physical and wireless communications links. The record link program 300 may be downloaded to the persistent storage 308 through the communications unit 410.

[0063] The I / O interface 412 allows for input and output of data to and from other devices that may be connected to the computing system 400. For example, the I / O interface 412 may provide a connection to an external device 418, such as a keyboard, a keypad, a touch screen, or any other suitable input device or combination thereof. The external device 418 may also include portable computer-readable storage media, such as, for example, thumb drives, portable optical or magnetic disks, and memory cards. Software and data used to implement embodiments of the present invention, such as, for example, the record link program 300, may be stored on such portable computer-readable storage media and loaded into the persistent storage 408 via the I / O interface 412. The I / O interface 412 also connects to a display 420.

[0064] Display 420 provides a mechanism for displaying data to a user and may be, for example, a computer monitor.

[0065] The programs described herein are identified based on the applications for which they are implemented in particular embodiments of the invention. However, it should be understood that any specific program names herein are used for convenience only, and thus the invention should not be limited to use only with any particular application identified or suggested by such names, or combinations thereof.

[0066] The present invention may be a system, method or computer program product, or combination thereof, integrated at any possible level of technical detail. The computer program product may include a computer-readable storage medium having one (or more) computer-readable program instructions for causing a processor to perform aspects of the present invention.

[0067] A computer-readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of computer-readable storage media includes portable computer diskettes, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), static random access memories (SRAMs), portable compact disk read-only memories (CD-ROMs), digital versatile disks (DVDs), memory sticks, floppy disks, mechanically encoded devices such as punch cards or ridge-in-groove structures with recorded instructions, and any suitable combination of the foregoing. A computer-readable storage medium as used herein should not be construed as a transitory signal per se, such as electric waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., light pulses passing through a fiber optic cable), or electrical signals transmitted through wires.

[0068] The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to each computing / processing device, or may be downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, or a wireless network, or a combination thereof. The network may comprise copper transmission cables, optical transmission fiber, wireless transmission, routers, firewalls, switches, gateway computers, or edge servers, or a combination thereof. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and transfers the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing / processing device.

[0069] The computer readable program instructions for carrying out the operations of the present invention may be either assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, configuration data for an integrated circuit, or source or object code written in any combination of one or more programming languages, including object oriented programming languages ​​such as Smalltalk® or C++, procedural programming languages ​​such as the “C” programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), and the connection may be made to an external computer (e.g., through the Internet using an Internet Service Provider). In some embodiments, to carry out aspects of the invention, electronic circuitry including, for example, a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may execute computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry.

[0070] Aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations or block diagrams or combinations thereof, and combinations of each block in the flowchart illustrations or block diagrams or combinations thereof, can be implemented by computer readable program instructions.

[0071] These computer readable program instructions may be provided to a computer processor or other programmable data processing apparatus producing a machine, such that the instructions executed via the computer processor or other programmable data processing apparatus create means for implementing the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams. These computer readable program instructions may also be stored on a computer readable storage medium capable of instructing a computer, programmable data processing apparatus or other device, or combination thereof, to function in a particular manner, such that a computer readable storage medium having instructions stored thereon comprises an article of manufacture including instructions that implement aspects of the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams.

[0072] Also, the computer readable program instructions may be loaded into a computer, other programmable data processing apparatus, or other device to cause a sequence of operational steps to be executed on the computer, other programmable apparatus, or other device to generate a computer-implemented process, such that the instructions executing on the computer, other programmable apparatus, or other device implement the functions / operations specified in one or more blocks of the flowcharts and / or block diagrams.

[0073] The flowcharts and block diagrams in the drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowcharts or block diagrams may represent a module, segment, or portion of instructions that includes one or more executable instructions that implement a specified logical function. In some alternative implementations, the functions described in the blocks may be performed out of the order described in the figures. For example, two blocks shown in succession may actually be realized as one step and executed simultaneously or substantially simultaneously in a partially or fully overlapping manner in time, and in some cases, these blocks may be executed in reverse order depending on the functions involved. It should also be noted that each block of the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, may be implemented by a dedicated hardware-based system that performs the specified functions or operations or executes a combination of dedicated hardware and computer instructions. According to this specification, the following items are also disclosed. (Item 1) 1. A method for resolving entity records in a master data management (MDM) system, comprising: receiving, by one or more processors, a first score associated with a match between respective entity records of a plurality of entities of the master data of the MDM system; the one or more processors identifying a set of unresolved entity records, wherein the first score between pairings of respective entity records in the set of unresolved entity records is above a lower threshold score and less than an upper threshold score; generating, by the one or more processors, a second score associated with a relationship between the pairing of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system; generating, by the one or more processors, an overall score by combining the first score and the second score for the pairing of the respective entity records in the set of unresolved entity records; the one or more processors determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the aggregate score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, the one or more processors combine information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; A method comprising: (Item 2) 2. The method of claim 1, wherein the first score is generated excluding relationship information, hierarchy information and grouping information of the respective entity records of the plurality of entities of the master data of the MDM system. (Item 3) 3. The method of claim 1 or 2, wherein the second score is based on additional relationship information, grouping and hierarchy information associated with entities of the MDM system. (Item 4) the one or more processors performing, for each entity of the plurality of entities, a matching assessment between a first entity of the plurality of entities of the MDM system and a second entity of the plurality of entities; the one or more processors generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; 4. The method of any one of items 1 to 3, further comprising: (Item 5) 5. The method of any one of claims 1 to 4, wherein the second score is based on qualified data of relationship information, hierarchy information and grouping information associated with the pairing of each of the unresolved entity records and includes a weighting factor for determining that a relationship exists for the pairing of each of the unresolved entity records with a third entity and a weighting factor for determining that a relationship does not exist. (Item 6) removing redundant entity records from the master data of the MDM system, in response to the one or more processors combining information of the pairing of the respective entity records of the set of unresolved entity records into the single entity record. 6. The method of any one of items 1 to 5, further comprising: (Item 7) creating, by the one or more processors, a machine learning model that generates, based on the relationship data of the plurality of entities, the second score associated with the relationship between each of the pairings of the set of unresolved entity records; receiving, by the one or more processors, second scores and weighting factors corresponding to relationship types, hierarchy conditions and common grouping attributes of the set of unresolved entity records; training the machine learning model by applying, by the one or more processors, the second scores and the weighting factors corresponding to the relationship types, hierarchy conditions and common grouping attributes of the set of unresolved entity records as supervised learning; applying, by the one or more processors, a machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records; 7. The method of any one of items 1 to 6, further comprising: (Item 8) 1. A computer program product for resolving entity records in a master data management (MDM) system, the computer program product comprising: receiving a first score associated with a match between each entity record of a plurality of entities in the master data of the MDM system; identifying a set of unresolved entity records, wherein the first score between each pairing of entity records in the set of unresolved entity records is above a lower threshold score and below an upper threshold score; generating a second score associated with a relationship between the pairing of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system; generating a total score by combining the first score and the second score for the pairing of each entity record in the set of unresolved entity records; determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the aggregate score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, combining information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; having Computer program. (Item 9) 9. The computer program of claim 8, wherein the first score is generated excluding relationship information, hierarchy information, and grouping information of the entity records of each of the plurality of entities of the master data of the MDM system. (Item 10) 10. The computer program of claim 8 or 9, wherein the second score is based on additional relationship information, grouping and hierarchy information associated with entities of the MDM system. (Item 11) The processor, for each entity of the plurality of entities, performing a matching assessment of the MDM system between a first entity of the plurality of entities and a second entity of the plurality of entities; generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; 11. The computer program according to any one of items 8 to 10, further comprising: (Item 12) 12. The computer program product of claim 8, wherein the second score is based on relationship, hierarchy and grouping information associated with the pairing of each of the unresolved entity records and includes a weighting factor for determining that a relationship exists for the pairing of each of the unresolved entity records with a third entity and a weighting factor for determining that a relationship does not exist. (Item 13) The processor, removing redundant entity records from the master data of the MDM system in response to combining the pairing information of the respective entity records of the set of unresolved entity records into the single entity record. 13. The computer program according to any one of items 8 to 12, further comprising: (Item 14) The processor, creating a machine learning model that generates the second score associated with the relationship between each of the pairings of the set of unresolved entity records based on the relationship data of the plurality of entities; receiving second scores and weighting factors corresponding to relationship types, hierarchy conditions and common grouping attributes of the set of unresolved entity records; training the machine learning model by applying the second scores and the weighting factors corresponding to the relationship types, hierarchy conditions and common grouping attributes of the set of unresolved entity records as supervised learning; applying the machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records; 14. The computer program according to any one of items 8 to 13, further comprising: (Item 15) 1. A computer system for resolving entity records in a master data management (MDM) system, comprising: one or more computer processors; one or more computer readable storage media; program instructions stored on said one or more computer readable storage media; Equipped with The program instructions include: program instructions for receiving a first score associated with a match between each entity record of a plurality of entities of the master data of the MDM system; program instructions for identifying a set of unresolved entity records, wherein the first score between pairings of respective entity records in the set of unresolved entity records is greater than a lower threshold score and less than an upper threshold score; program instructions for generating a second score associated with a relationship between the pairing of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system; program instructions for generating a total score by combining the first score and the second score for the pairing of each entity record in the set of unresolved entity records; program instructions for determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the total score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, program instructions for combining information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; having Computer system. (Item 16) 16. The computer system of claim 15, wherein the first score is generated excluding relationship information, hierarchy information and grouping information of the respective entity records of the plurality of entities of the master data of the MDM system, and the second score is based on additional relationship information, hierarchy information and grouping information associated with the respective entity records of the plurality of entities of the master data of the MDM system. (Item 17) program instructions for performing, for each entity of the plurality of entities, a matching assessment between a first entity of the plurality of entities and a second entity of the plurality of entities of the MDM system; generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; 17. The computer system of item 15 or 16, further comprising: (Item 18) 18. The computer system of claim 15, wherein the second score is based on relationship information, hierarchy information, and grouping information associated with the pairing of each of the unresolved entity records and includes a weighting factor for determining that a relationship exists for the pairing of each of the unresolved entity records with a third entity and a weighting factor for determining that a relationship does not exist. (Item 19) program instructions for removing redundant entity records from the master data of the MDM system in response to combining information of the pairing of the respective entity records of the set of unresolved entity records into the single entity record; 19. The computer system of any one of items 15 to 18, further comprising: (Item 20) program instructions for creating a machine learning model that generates the second score associated with the relationship between each of the pairings of the set of unresolved entity records based on the relationship data of the plurality of entities; program instructions for receiving second scores and weighting factors corresponding to relationship types, hierarchy conditions, and common grouping attributes of the set of unresolved entity records; program instructions for training the machine learning model by applying the second scores and the weighting factors corresponding to the relationship types, the hierarchy conditions, and the common grouping attributes of the set of unresolved entity records as supervised learning; program instructions for applying a machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records; 20. The computer system of any one of items 15 to 19, further comprising:

Claims

1. 1. A method for resolving redundant entity records in a master data management (MDM) system, comprising: receiving, by one or more processors, a first score associated with a match between pairings of entity records of respective entities of a plurality of entities of a master data of an MDM system, the first score being generated by a comparison of attributes between the pairings of entity records; the one or more processors identifying a set of unresolved entity records, wherein the first score associated with a match between the pairings of each entity record in the set of unresolved entity records is above a lower threshold score and less than an upper threshold score; generating, by the one or more processors, a second score associated with a relationship between the pairings of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system, the second score being based on relationships, common groupings, and hierarchical structures between the pairings of the respective entity records and with other entity records included in the master data of the MDM system; the one or more processors generating a total score for the pairing of each entity record in the set of unresolved entity records, the total score being a sum of the first score and the second score; the one or more processors determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the total score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, the one or more processors combine information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; A method comprising:

2. 2. The method of claim 1 , wherein the first score is generated excluding the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records of the plurality of entities of the master data of the MDM system and with other entity records included in the master data of the MDM system.

3. 3. The method of claim 1 or 2, wherein the second score is based on the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records associated with an entity of the MDM system and with other entity records included in the master data of the MDM system.

4. The first score is generated by comparing attributes between pairings of entity records, the step comprising: the one or more processors performing, for each entity of the plurality of entities, a matching evaluation by comparing attributes of a first entity of the plurality of entities and a second entity of the plurality of entities of the MDM system; the one or more processors generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; The method according to any one of claims 1 to 3, comprising:

5. 5. The method of claim 1, wherein the second score is based on the relationships, common groupings, and hierarchical structures associated with the relationships between the pairings of the respective entity records of the unresolved entity records, and includes a weighting factor for determining that a relationship exists between the pairings of the respective entity records of the unresolved entity records and other common entities included in the master data of the MDM system, and a weighting factor for determining that a relationship does not exist between the pairings of the respective entity records of the unresolved entity records and other entities included in the master data of the MDM system.

6. the one or more processors, in response to combining information of the pairings of the respective entity records of the set of unresolved entity records to the single entity record, removing redundant entity records from the master data of the MDM system, the redundant entity records being entity records other than the single entity record for which the total score associated with the pairings of the respective entity records of the set of unresolved entity records exceeds the upper threshold score. The method of claim 1 , further comprising:

7. creating, by the one or more processors, a machine learning model that generates, based on the relationship data of the plurality of entities, the second score associated with the relationship between each of the pairings of the set of unresolved entity records; receiving, by the one or more processors, second scores and weighting factors corresponding to relationships, common groupings, and hierarchical structures of the set of unresolved entity records; and training the machine learning model by applying, by the one or more processors, the second scores and the weighting factors corresponding to the relationships, common groupings, and hierarchical structures of the set of unresolved entity records as supervised learning; The step of generating the second score comprises: applying, by the one or more processors, a machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records to generate the second scores. The method of any one of claims 1 to 6, comprising:

8. 1. A computer program product for resolving redundant entity records in a master data management (MDM) system, the computer program product comprising: receiving a first score associated with a match between a pairing of entity records of each of a plurality of entities of master data of the MDM system, the first score being generated by a comparison of attributes between the pairing of entity records; identifying a set of unresolved entity records, wherein the first score associated with a match between the pairings of each entity record in the set of unresolved entity records is greater than a lower threshold score and less than an upper threshold score; generating a second score associated with a relationship between the pairings of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system, the second score being based on relationships, common groupings, and hierarchical structures between the pairings of the respective entity records and with other entity records included in the master data of the MDM system; generating a total score for the pairing of each entity record in the set of unresolved entity records, the total score being a sum of the first score and the second score; determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the aggregate score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, combining information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; having Computer program.

9. 9. The computer program product of claim 8, wherein the first score is generated excluding the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records of the plurality of entities of the master data of the MDM system and with other entity records included in the master data of the MDM system.

10. 10. The computer program product of claim 8 or 9, wherein the second score is based on the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records associated with an entity of the MDM system and with other entity records included in the master data of the MDM system.

11. The first score is generated by comparing attributes between a pairing of entity records, the procedure comprising: The processor, performing, for each entity of the plurality of entities, a matching evaluation by comparing attributes of a first entity of the plurality of entities with a second entity of the plurality of entities of the MDM system; generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; A computer program according to any one of claims 8 to 10, comprising:

12. 12. The computer program product of claim 8, wherein the second score is based on the relationships, common groupings, and hierarchical structures associated with the relationships between the pairings of the respective entity records of the unresolved entity records, and includes a weighting factor for determining that a relationship exists between the pairings of the respective entity records of the unresolved entity records and other common entities included in the master data of the MDM system, and a weighting factor for determining that a relationship does not exist between the pairings of the respective entity records of the unresolved entity records and other entities included in the master data of the MDM system.

13. The processor, removing redundant entity records from the master data of the MDM system in response to combining information of the pairings of the respective entity records of the set of unresolved entity records to the single entity record, the redundant entity records being entity records other than the single entity record for which the total score associated with the pairings of the respective entity records of the set of unresolved entity records exceeds the upper threshold score; 13. A computer program product as claimed in any one of claims 8 to 12, further comprising:

14. The processor, creating a machine learning model that generates the second score associated with the relationship between the pairings of each of the set of unresolved entity records based on the relationship data of the plurality of entities; receiving second scores and weighting factors corresponding to relationships, common groupings, and hierarchical structures of the set of unresolved entity records; training the machine learning model by applying the second scores and the weighting factors corresponding to the relationships, common groupings, and hierarchical structures of the set of unresolved entity records as supervised learning; The step of generating the second score includes: applying the machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records to generate the second scores.

14. A computer program according to any one of claims 8 to 13, comprising:

15. 1. A computer system for resolving redundant entity records in a master data management (MDM) system, comprising: one or more computer processors; one or more computer readable storage media; program instructions stored on said one or more computer readable storage media; Equipped with The program instructions include: program instructions for receiving a first score associated with a match between a pairing of entity records of each of a plurality of entities of a master data of an MDM system, the first score being generated by a comparison of attributes between the pairing of entity records; program instructions for identifying a set of unresolved entity records, wherein the first score associated with a match between the pairing of each entity record in the set of unresolved entity records is greater than a lower threshold score and less than an upper threshold score; program instructions for generating a second score associated with a relationship between the pairings of the respective entity records of the unresolved entity records based on relationship data of the plurality of entities added to the master data of the MDM system, the second score being based on relationships, common groupings, and hierarchical structures between the pairings of the respective entity records and with other entity records included in the master data of the MDM system; program instructions for generating a total score, the total score being a sum of the first score and the second score, for the pairing of each entity record in the set of unresolved entity records; program instructions for determining whether the total score associated with the pairing of the respective entity records in the set of unresolved entity records exceeds the upper threshold score; in response to the total score of the pairings of the respective entity records in the set of unresolved entity records exceeding the upper threshold score, program instructions for combining information of the pairings of the respective entity records in the set of unresolved entity records into a single entity record; having Computer system.

16. 16. The computer system of claim 15, wherein the first score is generated excluding the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records of the plurality of entities of the master data of the MDM system and with other entity records included in the master data of the MDM system, and the second score is based on the relationships, including the relationships, common groupings, and hierarchical structures, between the pairings of the respective entity records of the plurality of entities of the master data of the MDM system and with other entity records included in the master data of the MDM system.

17. The first score is generated by comparing attributes between a pairing of entity records, the program instructions comprising: program instructions for performing, for each entity of the plurality of entities, a matching evaluation by comparing attributes of a first entity of the plurality of entities to a second entity of the plurality of entities of the MDM system; generating the first score associated with a match between the first entity of the plurality of entities and the second entity of the plurality of entities; 17. A computer system according to claim 15 or 16, comprising:

18. 18. The computer system of claim 15, wherein the second score is based on the relationships, common groupings, and hierarchical structures associated with the relationships between the pairings of the respective entity records of the unresolved entity records, and includes a weighting factor for determining that a relationship exists between the pairings of the respective entity records of the unresolved entity records and other common entities included in the master data of the MDM system, and a weighting factor for determining that a relationship does not exist between the pairings of the respective entity records of the unresolved entity records and other entities included in the master data of the MDM system.

19. program instructions for removing redundant entity records from the master data of the MDM system in response to combining information of the pairings of the respective entity records of the set of unresolved entity records into the single entity record, the redundant entity records being entity records other than the single entity record for which the total score associated with the pairings of the respective entity records of the set of unresolved entity records exceeds the upper threshold score; 19. The computer system of claim 15, further comprising:

20. program instructions for creating a machine learning model that generates the second score associated with the relationship between the pairings of each of the set of unresolved entity records based on the relationship data of the plurality of entities; program instructions for receiving second scores and weighting factors corresponding to relationships, common groupings, and hierarchical structures of the set of unresolved entity records; and program instructions for training the machine learning model by applying the second scores and the weighting factors corresponding to the relationships, common groupings, and hierarchical structures of the set of unresolved entity records as supervised learning; The program instructions for generating the second score include: program instructions for applying a machine learning model trained with the second scores of the set of unresolved entity records and the weighting factors to a new set of unresolved entity records to generate the second scores.

20. A computer system according to any one of claims 15 to 19, comprising:

Images