searchOrganizationProbabilistic

Probabilistic searches use statistical methods to determine how closely party records match the given set of search criteria. The probabilistic search engine generates match scores that consider the frequency of the occurrence of a given data value within a particular distribution.

For example, in English speaking locales, the organization name Johnson Ltd. may contribute less to a match score than the name Devlin-Baker Ltd. because Johnson is more common.

By contrast, the classic InfoSphere® MDM deterministic search, as used by searchOrganization, derives its match scores by determining if there is an exact match between the search criteria and the corresponding data in a candidate record.

In probabilistic searches, each criterion provided in the search request will contribute to the search score for each record being searched. The search results will return only records that meet a minimum search score threshold.

When using the default probabilistic matching algorithm, one or more of the following search criteria can be provided:

Organization Name

• TCRMOrganizationNameBObj.NameUsageType
• TCRMOrganizationNameBObj.NameUsageValue
• TCRMOrganizationNameBObj.OrganizationName

Address

• TCRMPartyAddressBObj.AddressUsageType
• TCRMPartyAddressBObj.AddressUsageValue
• TCRMAddressBObj.AddressLineOne
• TCRMAddressBObj.AddressLineTwo
• TCRMAddressBObj.City
• TCRMAddressBObj.ResidenceNumber
• TCRMAddressBObj.ProvinceStateType
• TCRMAddressBObj.ProvinceStateValue
• TCRMAddressBObj.ZipPostalCode

The default searchOrganizationProbabilistic transaction only considers the following address usage types:

• Business address
• Mailing address

The AddressUsageType and AddressUsageValue are optional. The search transaction compares the provided address against the addresses of all usage types in the InfoSphere MDM Probabilistic Matching Engine indices.

Party Identification

• TCRMPartyIdentificationBObj.IdentificationType
• TCRMPartyIdentificationBObj.IdentificationValue
• TCRMPartyIdentificationBObj.IdentificationNumber

The search transaction considers only Corporate Tax ID and DUNS Number.

The IdentificationType or IdentificationValue must be provided along with the IdentificationNumber.

Contact Method

• TCRMPartyContactMethodBObj.ContactMethodUsageType
• TCRMPartyContactMethodBObj.ContactMethodUsageValue
• TCRMContactMethodBObj.ReferenceNumber

The search transaction considers only the Business Telephone contact method usage type.

The ContactMethodUsageType and ContactMethodUsageValue are optional. If neither the ContactMethodUsageType nor ContactMethodUsageValue are provided, the transaction treats the phone number as a business telephone number.

The transaction request supports multiple organization names, addresses, contact methods, and identifications as search criteria. In general, when multiple business objects in a search are of the same type, they are searched with an OR relationship. Business objects of different types are searched with an AND relationship. For example:

• If two different names, Name1 and Name2, are provided as search criteria, then the search results correspond to candidates that match either Name1 or Name2.
• If a name, Name1, and a home address, Address1, are provided as search criteria, then the search results correspond to candidates that match both Name1 and Address1.
• If two names, Name1 and Name2, and a home address, Address1, are provided as search criteria, then the search results correspond to candidates that match either of the following:
  • Name1 and Address1
  • Name2 and Address1

Other party data that are included in a search request but are not identified as a matching data element in the probabilistic matching algorithm will be ignored by the transaction.

To override configuration defaults, you can provide the following additional parameters in the search request:

MaxReturn

The maximum number of party records that the transaction will return.

By default, the maximum number of party records returned in the transaction response is configurable in the Configuration and Management component. However, you can use the MaxReturn parameter in the transaction request to override the configured default maximum.

Restriction: You can only use MaxReturn to set the maximum number of returned parties to be less than the configured value, not larger. If the MaxReturn value is higher than the configured value, then the number of parties returned in the response will be set to the configured maximum value.

MinSuspectType

The minimum suspect type threshold for which matched records will be returned.

This value can be provided to override the default suspect type of a B match (a possible match). Valid values for MinSuspectType are defined in the CDSUSPECTTP code table:

• A value of 1 indicates a probabilistic match threshold type of Match.
• A value of 3 indicates a probabilistic match threshold type of Possible Match.
• A value of 4 indicates a probabilistic match threshold type of No Match.

MinScore

The minimum score for which matched parties will be returned. This numeric value can be provided to override the default setting of a possible match (a B match). If the MinSuspectType is also provided in the request, then the transaction will use the MinSuspectType value instead of the MinScore value.

This transaction supports the Pagination feature.

• The InfoSphere MDM Probabilistic Matching Engine must be configured.
• Party matching data must be synchronized with the InfoSphere MDM Probabilistic Matching Engine indices.

• TCRMOrganizationNameBObj.OrganizationName
• TCRMAddressBObj.AddressLineOne
• TCRMAddressBObj.AddressLineTwo
• TCRMAddressBObj.City
• TCRMAddressBObj.ResidenceNumber
• TCRMAddressBObj.ProvinceStateType
• TCRMAddressBObj.ProvinceStateValue
• TCRMAddressBObj.ZipPostalCode
• TCRMContactMethodBObj.ReferenceNumber
• TCRMPartyIdentificationBObj.IdentificationNumber
  Note: If the TCRMPartyIdentificationBObj.IdentificationNumber is provided, you must also include one of the following:

  • TCRMPartyIdentificationBObj.IdentificationType
  • TCRMPartyIdentificationBObj.IdentificationValue

The search transaction uses the search criteria to identify a list of candidates through the InfoSphere MDM Probabilistic Matching Engine bucketing strategy. There are several default search buckets configured.

Name

Nicknames and phonetic values, as defined in the InfoSphere MDM Probabilistic Matching Engine configuration, are considered.

Name and ZIP/postal code

There is an exact match on the ZIP or postal code and, for name, any one or more of the name tokens.

Both InfoSphere MDM Probabilistic Matching Engine-defined nicknames and phonetic values, as defined in the configuration, are considered for the name.

Business telephone

There is an exact match on the Business Telephone type of contact method.

Corporate Tax ID

There is an exact match on a Corporate Tax ID type of identification number.

DUNS Number

There is an exact match on a DUNS Number type of identification number.

After all candidates have been identified using the bucketing strategy, each candidate party is compared against all of the search criteria and assigned a search score that represents how closely the candidate matches the search criteria. The search transaction only returns records that have a score higher than a given search threshold.

The InfoSphere MDM Probabilistic Matching Engine indices capture matching data for active parties only so, in general, the transaction can only return parties that are active.

For the matched records returned by the search, the types of party data returned for address, name, and party identification are governed by default search configurations. The transaction returns inactive names only if no active names are available. However, only active values for address, contact number, and party identification are returned.

For example, if the search return value for addresses (/IBM/Party/Search/ReturnValue/organizationAddressUsageType) is set to 1, then the first active address found with a usage type of 1 is returned in the search results. If the address with a usage type of 1 is inactive, then the transaction returns the next available active address associated with the party instead. If there is no active address, then no address is returned.

<TCRMTxObject> ProbabilisticOrganizationSearchBObj

<TCRMObject> ProbabilisticOrganizationSearchBObj

with one or more business objects:

• TCRMOrganizationBObj
• TCRMOrganizationNameBObj
• TCRMPartyAddressBObj, each containing TCRMAddressBObj
• TCRMPartyContactMethodBObj, each containing TCRMContactMethodBObj
• TCRMPartyIdentificationBObj