公开(公告)号：GB2504966A

公开(公告)日：2014-02-19

申请号：GB201214542

申请日：2012-08-15

Applicant: IBM

Inventor： BAUDEL THOMAS ,  BROEKSEMA ALBERTUS HENDRIK JOHAN

IPC: G06F17/30 ,  G06T11/20

Abstract: A method of processing a data plot comprising a plurality of data points for inclusion of additional information content is presented. The method comprises the steps of: dividing a space of the data plot into subspaces, each subspace containing at least one data point of the data plot; computing an available area for each subspace; and computing a compressed information representation for each subspace based on information about the least one data point contained in the subspace and the computed available area for the subspace. Preferably the space is divided according to Voronoi decomposition and if the subspaces are below a threshold size then they are merged with neighbouring subspaces.

公开(公告)号：GB2505220A

公开(公告)日：2014-02-26

申请号：GB201215021

申请日：2012-08-23

Applicant: IBM

Inventor： FEILLET PIERRE DENIS ,  BAUDEL THOMAS

IPC: G06Q10/06 ,  G06F17/30 ,  G06N5/02 ,  G06Q40/08

Abstract: A method, system and computer program are provided for rule execution in decision services, such as an insurance company. The method includes providing one or more specialized (i.e. reduced) rulesets classified by decision partitions based on clustering predicates of decision service signatures for business rules. A decision request is received 201 from a client and one or more parameters of the decision request are input 203A to a clustering predicate based on decision service signatures to produce a decision category. A specialized ruleset for the decision request is determined 203B based on the decision category, and the decision request is forwarded 205 for processing by the specialized ruleset. A decision output is returned 207 from the specialized ruleset. The provision of the specialized rulesets may include decomposing a monolithic multipurpose ruleset. The method may also include defining a clustering predicate at business rules authoring time, and analysing business rules to identify decision partitions based on the clustering predicate and to classify the business rules per decision partition.

公开(公告)号：GB2509538A

公开(公告)日：2014-07-09

申请号：GB201300256

申请日：2013-01-08

Applicant: IBM

Inventor： SAUTEREY NICOLAS ,  BAUDEL THOMAS ,  ZENTNER CLAUDIA

IPC: G06Q10/06

Abstract: A method for process instance migration from a former process model to a new process model comprises analysing a former process model and a new process model and their data models to extract a process change model relating to changes in the new process model on which to define migration rules, analysing one or more existing process instances and retaining process instances whose current state has passed the changes in the new process model, generating candidate activities of the retained process instances, generating a set of conditions for the candidate activities, and generating a set of actions to be taken if a condition is met. The method may also include presenting the generated set of conditions and set of actions to a user enabling the user to combine conditions and actions to produce a process instance migration program in the form of a set of rules. A system and computer program product for implementing the method are also claimed. The invention is intended to ensure consistency when migrating running process instances form an old version of a business process to a new version.

公开(公告)号：CA2783451C

公开(公告)日：2018-01-16

申请号：CA2783451

申请日：2011-04-21

Applicant: IBM

Inventor： DELHOUME FREDERIC ,  BAUDEL THOMAS

IPC: G06F9/451

Abstract: The invention provides a method for generating a display of a source data model on a user interface component being associated with a target data model. The source data model and the target data model comprise a collection of data types (class/object) each comprising a set of named and types attributes. The method comprises : a. for each source data type of the source data model determining a matching target type among said target types in the target data model, b. linking said source type to said matching target type, c. generating a display of said source data model using said link between a source type and a target type. Step a comprises: for each target type in the target data model, determining an attribute score matrix storing a score value for each correspondence between each target attribute in a first attribute set of said target attribute and each source attribute of said source data type, based on a set of predefined matching rules, determining a target type score for the target type based on the attribute score matrix, and determining the matching target type for said source type based on the target type scores associated with the target types of the target model.

公开(公告)号：CA2783451A1

公开(公告)日：2011-12-08

申请号：CA2783451

申请日：2011-04-21

Applicant: IBM

Inventor： DELHOUME FREDERIC ,  BAUDEL THOMAS

IPC: G06F9/451

Abstract: The invention provides a method for generating a display of a source data model on a user interface component being associated with a target data model. The source data model and the target data model comprise a collection of data types (class/object) each comprising a set of named and types attributes. The method comprises : a. for each source data type of the source data model determining a matching target type among said target types in the target data model, b. linking said source type to said matching target type, c. generating a display of said source data model using said link between a source type and a target type. Step a comprises: for each target type in the target data model, determining an attribute score matrix storing a score value for each correspondence between each target attribute in a first attribute set of said target attribute and each source attribute of said source data type, based on a set of predefined matching rules, determining a target type score for the target type based on the attribute score matrix, and determining the matching target type for said source type based on the target type scores associated with the target types of the target model.