Microsoft PowerPoint - Structured and Unstructured Information by leader6


									The Role of Structured & Unstructured
  Information in Decision Support
   Overview in the context of small to medium sized businesses

              Azam N Foda CMA, MBA

                               Small to Medium sized Business
• The field of Expert Systems & Artificial Intelligence is a well
  known researched field for organizations involved in areas such
  as robotics, space and sophisticated manufacturing with a need
  for remote diagnostics and automation.

• Small & medium sized organizations have found Expert Systems
  out of reach due to various reasons:
    – Magnitude of investment necessary.
    – The ability to understand and manage the systems.
    – The availability of supporting infrastructure.

This presentation attempts to highlight the role of unstructured information by
   relating existing concepts and knowledge to business decisions in small to
   medium sized organizations.

                   Definition: Decision Support Systems

Decision Support Systems (DSS) are interactive
 computer-based systems used by managers as an aid to
 decision making in semi-structured decision tasks… J
 Efrim Boritz (1991)

The impact is on decisions in which there is sufficient
  structure for computer and analytic aids to be of value
  but where manager’s judgment is essential…. Ralph
  Sprague (1980)

                  Traditional domain of DSS - Structured
• Herbert Simon (1960) suggested that problems could
  be classified into three main categories:
   – Structured.
   – Semi-Structured
   – Unstructured

• Structured problems are easier to solve algorithmically
  and have been the domain of DSS tools given that the
  underlying data is also structured.

                       Manifestations of Structured DSS in
• Ability to make inquiries through a computer terminal
  about the status of various aspects of a company, such
   –   Sales trends
   –   Production trends                             ed tion
                                                   er a
   –   Wage costs and trends                    ay rm
                                               L fo
   –   Output specifications, scrap, rework     In

• Ability to drill down through the information layers and
  determine further breakdown of summarized
                       Complementing Structured DSS
• Good Business Decisions can be made by complementing
  Structured DSS with Unstructured Information.

• Decisions made purely on structured information lack discretion
  and flexibility required to be successful in business.

• In order to better leverage unstructured information for decision
  making one needs to supplement numerical computations with
  the capability to use domain knowledge, reasoning and judgment.

                                Information Required for Business
Business decisions need structured as well as unstructured
  information, examples:

   – Credit and Collection
       • Flexibility based on volume of business both current and anticipated, customer
         relationships, input from last customer meeting….

   – Pricing and Quotes
       • Flexibility based on volume of business, market penetration, supplementary business to
         be derived…..

   – Employee Training
       • Discretion based on performance, succession plan, interpersonal skills……

   – Purchasing and vendor selection
       • Discretion based on service, reliability, knowledge of your industry, longevity…..

                              Availability of Unstructured
Unstructured Information is spread all over an
 organization & it’s people. In an organizational
 context, it includes:

     •   Research Reports
     •   Customer Meeting Notes
     •   Sales Collateral
     •   Development & Engineering Specifications
     •   Internal Memos
     •   Patents & Invention Proposals
     •   Press Releases and News Articles
     •   E-mail Messages

                         Knowledge Based Decision Support Systems
                                                    (KBDSS     )
Expert systems or Knowledge Based Decision Support
  Systems (KBDSS) combine structured and unstructured

   – Traditionally:
      • Data + Model                DSS
         – Model = a system of postulates, data, and inferences presented as a
           mathematical description of an entity or state of affairs. Merriam Webster

   – Now:
      • Data + Model + Knowledge                     KBDSS

                               KBDSS Characteristics
A KBDSS has domain knowledge and heuristic reasoning
  mechanisms incorporated in the system .. Feigenbaum 1981

   – The domain knowledge relates to “facts” that are widely shared
     and agreed upon by experts in the field.

   – The heuristic reasoning refers to mostly private and little discussed
     rules of judgment that characterize expert level decision making.
     This is derived from the reservoir of experience and is a function
     of the quality of knowledge and the experience accumulated over
     time in integrating the different pieces together.

                                                         Two Dimensional View of DSS

                                                                                                            TEM CE
                                                                                                         YS GEN
                                    Implicit Detail                                                R  T S LLI
                                                                                                 PE TE
                                                                                                X N
                                                                                             S/E AL I
    Computational Procedures

                                                                                         BD FICI
                                                                                        K TI

                                                                  Decision Support Systems
                                       General Problem Solver
                                       Transaction Processing

                                                                Database mgmt systems

                                         Programs and Files
                                                                                        Implicit Detail
                           e ta
                                                      Data Retrieval Procedures
Exp                                                                                              Adapted from Bonczek et al 1981
                       Capture and Preservation of unstructured
                       information in a Knowledge Base is key

• Organizations have a need to preserve their knowledge base
  because Human expertise is perishable at an organizational level
  due to:
   –   Transfer
   –   Retirement
   –   Resignation
   –   Rightsizing

• The exiting employees walk out the door with years of
  knowledge of specific relevance to the organization.

• There is much to be gained from capturing this knowledge in the
  enterprise reservoir for retrieval when necessary.
                             Leverage the existing information for
                                       Decision Making

Organizations need to build a capability to better leverage their
information for decision making.

     Dimension            Transaction         Decision Support     Knowledge based
                       Processing Systems         Systems          Decision Support
 Level of Problem          Clerical and          Strategic and       Technical and
 Solving                   Operational            Managerial          Professional
 Type of Problem           Repetitive         Complex, Broad and   Complex, Narrow
 Solving                                           ad hoc           and repetitive

 Primary Data              Numerical              Numerical            Symbolic
 Database Content         Record and          Raw and processed     Procedural and
                       Summarize raw data           data           Factual knowledge

                    STRUCTURED DATA              UNSTRUCTURED DATA

                                   Benefits of a KBDSS

The Benefits of a KBDSS to small and medium sized
  businesses are numerous, to name a few:

   –   Facilitates delegation without loss of comfort
   –   Increased breadth and thoroughness of information
   –   More defensible logic
   –   Separation of preferences from beliefs
   –   Removal of bias and introduction of consistency
   –   Preservation of enterprise expertise and knowledge
   –   Improved managerial effectiveness
   –   Desired level of flexibility to remain competitive in business

                                Analytical Hierarchy Process (AHP)
                                                          Thomas Saaty 1970

• A decision tool to help managers deal with unstructured
  problems which
   – have multiple and possibly conflicting objectives
   – require judgment and experience
   – combine qualitative and quantitative considerations
• Five steps in AHP process
   –   Define goal or objective
   –   Determine the criteria that bear on the problem
   –   Determine the alternative courses of action that achieve the goal
   –   Prioritize
        • The relative importance of each criteria
        • The relative strength of each alternative course of action
   – Select the best alternative or mix of alternatives

                                            A supplier decision problem in the
                                                       AHP context
Quality                   Product Quality             Quality Rating
                                                      Defect Record
                                                      Continuous Improvement Program
                          Quality Awareness           Quality of Prevention
                                                      Quality of Detection
                                                      Quality of Correction
                          Service Quality             Marketing Support
                                                      Manufacturing Support
                                                      Storage Facilities
Delivery                  Delivery Reliability
                          Delivery Time
                          Supplier Location
Cost                      Unit Price
                          Transportation Cost
Finance                   Supplier financial status
                          Supplier terms of payment
                          Supplier Costing System
Management                Style                     Communication Procedures
                                                    Management Commitment
                          Capabalities              Technical Expertise
                                                    Manufacturing Expertise               Know-how
                                                                                          Ability to handle special requests
                                                                                          Ability to meet deadlines
Source: Decision-Support and Expert Systems for M anagement Accountants-Howard Armitage and J Efrim Boritz

                               Capital Budgeting and Strategic
                                        Planning DSS
• According to J Tracy Rourke, CEO of Allen Bradley
    – Return on Investment, Internal rate of return-such abstractions focus on the time
      value of money and give little consideration to the strategic opportunities and threats
      presented by technological advances….The old investment models were developed
      to compare one project with another. They were not developed to be an aid to
      strategic thinking

                                                       Structured Information
    – Strategic planning and budgeting defines an integrative endeavor of planning and
      budgetary control. It should reflect management planning and company doctrine. It
      is not simply a process or procedure but is an outcome of analysis of all possible
      production, environmental and competitive factors involved

                                                  Unstructured Information and

                            Implications for Professionals-Food
                                        for thought
Knowledge gleaned over years of experience can be expressed and
  automated through a few hundred heuristics in a KBDSS

• In the context of the earlier examples
   – Will we need a credit and collections person?
   – What will be the role of a pricing and quotes analyst?
   – How will the function of a Purchasing manager change?

• Once automated
   – Will the automated function be outsourced to low cost countries?
   – How will the role of senior executives change?


To top