Do you want to find match member records, link member records, and perfect a search algorithm for your InfoSphere MDM Virtual implementation? Then this course is designed for you. The InfoSphere MDM Virtual Module Algorithms V.11 course prepares students to work with and customize the algorithm configurations deployed to the InfoSphere MDM Probabilistic Matching Engine (PME) for Virtual MDM implementations. The PME is the heart of all Matching, Linking, and Searching for entities (Person, Organization, etc) that exist in InfoSphere MDM. This course has a heavy emphasis on the exercises, where the students will implement the customization discussed in the course to perform matching, linking, and searching on fields not provided by the default implementation. At the end of this course, it is expected that students will feel comfortable customizing an algorithm for the PME for Virtual implementations.

• Understand how Matching and Linking work for both the Virtual Implementations of InfoSphere MDM
• Understand the MDM configuration project and database tables used by the PME
• Understand the PME Algorithms (Standardization, Bucketing and Comparison steps) and how to create and customize the algorithms using the workbench
• Understand how to analyze the Bucketing steps in an algorithm
• Understand how to generate weights for a given algorithm and how those weights are generated based on a sample database set
• Understand how to analyze the weights that are generated using the workbench
• Understand how to deploy the PME configuration for the Virtual implementations of InfoSphere MDM

This intermediate course is for Business and Technical Specialist working with the Matching, Linking, and Search services of InfoSphere MDM Virtual module.

It is recommended that you take the following course prior to enrolling in this course:

• (1Z801) or experience with InfoSphere MDM

PME and Virtual Overview

• Virtual MDM Overview
• Terminology (Source, Entity, Member, Attributes)
• PME and Virtual MDM ( Algorithms, Weights, Comparison Scores, Thresholds)
• Virtual MDM Linkages and Tasks

Virtual MDM Algorithms

• Standardization
• Bucketing
• Comparison Functions
• Exercise: Creating a new Algorithm

Virtual PME Data Model

• Algorithm configuration tables
• Member Derived Data
• Bucketing Data
• Exercise: Loading Members and viewing Algorithm and Derived data

Bucket Analysis

• Analysis Overview
• Attribute Completeness
• Bucket Analysis
• Exercise: Analyzing our Buckets

Weights

• Weights Overview (Frequency-based weights, Edit Distance weights and Parameterize weights)
• The weight formula
• Running weight generation
• Analyzing weights
• Bulk Cross Match process
• Pair Manager
• Threshold calculations
• Exercise: Generate Weights and analyzing weight distribution
• Exercise: Pair Manager and Threshold Calculations
• Exercise: Testing our algorithm

Agenda:

Day 1

• Unit 1: PME and Virtual Overview
• Unit 2: Virtual MDM Algorithms
• Unit 3: Virtual PME Data Model

Day 2

• Unit 4: Bucket Analysis
• Unit 5: Weights

Prior to enrolling, IBM Employees must follow their Division/Department processes to obtain approval to attend this public training class. Failure to follow Division/Department approval processes may result in the IBM Employee being personally responsible for the class charges.  

GBS practitioners that use the EViTA system for requesting external training should use that same process for this course. Go to the EViTA site to start this process: http://w3.ibm.com/services/gbs/evita/BCSVTEnrl.nsf

Once you enroll in a GTP class, you will receive a confirmation letter that should show:

• The current GTP list price
• The 20% discounted price available to IBMers. This is the price you will be invoiced for the class.