Paper presentation

1
A Framework for Design Phase Prediction using
Integrated Product and Process Attribute approach

• Paper presentation
• by
• Prof. A. K. Verma
• Mr. Piyush Mehta
• Prof. A. Srividya

Indian Institute of Technology, Mumbai
November 2005
2
Model Premise
3
Measures Requirements Analysis End
Requirements Analysis Phase End Requirements Analysis Phase End Requirements Analysis Phase End Requirements Analysis Phase End Requirements Analysis Phase End Requirements Analysis Phase End
Engineering Thread Engineering Thread Engineering Thread Defect Detection Thread Defect Detection Thread Defect Detection Thread
Product Process Derived Product Process Derived
Specification Size Total Effort Productivity Review Coverage Review Effort Defect Density
Specification Complexity Total Duration Effort Variance Review Size Review Duration Review Efficiency
Specification Defects Total Team Size Size Variance Review Team Size Defect Variance
Team Competency Duration Variance Review Competency
Rework Effort
4
Design Effort Prediction Models
Multiple Regression Model I Dependent Variable
Design Effort Independent Variables
Specification Size, Specification Defects,
Specification Effort Factor Variable Complexity
Design Effort f (Specification Size,
Specification Complexity, Specification Defects,
Specification Effort) Multiple Regression Model
II Dependent Variable Design Effort
Independent Variables Specification Size,
Specification Defects, Specification
Effort) Design Effort f (Specification Size,
Specification Defects, Specification
Effort) Assumption Team competency is roughly
same and is taken into account while calculating
Specification Size Artificial Neural Network
Model for I and II
5
Design Duration Prediction Models
Multiple Regression Model for design duration
prediction Design Duration f (Specification
Size, Specification Effort, Specification
Duration, Specification Team Size, Potential
Design Team Size) Assumption Design Team Size
is constraint in the organization and depends on
the availability of the resources in the
organization Neural Network Model for design
duration prediction Design Duration f
(Specification Size, Specification Effort,
Specification Duration, Specification Team Size,
Design Team Size)
6
Design Size Prediction Models
Multiple Regression Model for design size
prediction Design Size f (Specification Size,
Specification Complexity, Specification
Effort) Assumption The model will be fitted on
the existing project database with the given
specification size, specification complexity,
specification effort. Specification size has
strong relationship with the design size. The
complexity drives the design size highly complex
requirements will correspond to higher design
size. Neural Network Model for design size
prediction Design Size f (Specification Size,
Specification Complexity, Specification Effort)
7
Defect Prediction Models
Multiple regression model for design defect
prediction Design Defect f (Design Size,
Design Complexity, Design Effort, Specification
Defects Density, Team Competency Index, Review
Team Size, Review Team Competency, Review
Effort) Assumption It can be observed to see
if the team size and team competency are
impacting each other. Parameter pruning can be
enabled to find out the critical parameters
important for predication of design defects.
Design size and review effort represents the
review pace, instead of review effort the review
pace can be used. Artificial Neural Network
model for design defect prediction Design Defect
f (Design Size, Design Complexity, Design
Effort, Specification Defects Density, Team
Competency Index, Review Team Size, Review Team
Competency, Review Effort)
8
Integrated Product and Process Attribute
Quantitative Model
9
Benefits of IPPA-QM Model

• Early prediction leads to better project planning
  
• Better project management and early corrective
  action (in case of deviation)
• Better decision making capability
• Understanding of organizational delivery
  capability
• Development of project at lower cost, better
  quality, at an agreed time and schedule with
  higher customer satisfaction.