Title: Microelectronics
1A Solution to the Recall Problem using Rough Set
Theory
Professor Djamel Bouchaffra
(Advisor) Tarek
Dakhlallah (Ph.D. Student)
Computer Science Engineering 131 Dodge Hall
Phone
248-370-2242, email bouchaff_at_oakland.edu
Mass production recall solutions
Current process
Current system inputs
- Recalls are necessary for a number of reasons,
today's heavy industries products are incredibly
complex with many interacting electronic systems - The development cycles are extremely compressed
- With such a short gestation period, certain
interactions or problems can go undetected in the
development process of a new vehicle
Manufacturer reports a problem
Product recall issued!
Customer reports a problem
Current process could be costly!
- Rough Set Theory has shown its fruitfulness in a
variety of data mining areas (information
retrieval, decision support, machine learning,
and knowledge based systems) - Concept is proven in may fields like medical data
analysis - and aircraft pilot performance evaluation
- An important result from the theory is that it
simplifies - the search for dominating attributes leading to
specific properties, or just rules pending in the
data
- Rough set theory proposes a formal framework for
the automated transformation of data into
knowledge - This method shows that the principles for
learning by examples can be formulated in the
basis of this theory - An important result from the theory is that it
simplifies the search for dominating attributes
leading to specific properties, or just rules
pending in the data
Proposed concept
Proposed concept
Rough Set Theory box
Proposed process
Vehicle is viewed as The system Information
source ltU,Agt
universe of vehicle components U
I can predict a possible future problem before
mass production
Current system inputs
Any collection X of vehicle components can be
estimated via the equivalence relation through
lower and upper approximations
Lower approximation Filters out noise of set X
and keeps only objects that matter the most
A is a finite set of attributes
Rough Set Theory Filter
Product recall issued!. But on early stage
Manufacturers product issues (recalls) history
Higher approximation
Objects that surely belong to the rough set
X. These vehicle components are of interest
Objects that approximately belong To rough set
Proposed process could be saving !