CSCI 781: Advanced Pattern Classification

Go back to CSCI 780

Go on to CSCI 871

Syllabus Date:

November, 1986

Advanced Pattern Classification. (3) (Prereq: CSCI 780) Feature nomination, selection, extraction, and evaluation; deterministic, stochastic, and fuzzy models for classifier design; error rate estimation; advanced architecture and processor design.

K.S. Fu, Syntactic Pattern Recognition and Applications, Prentice-Hall, Englewood Cliffs, 1982. (ISBN 0-13-880120-7)

R.C. Gonzalez, and M.G. Thomason, Pattern Recognition; An Introduction, Addison-Wesley, Reading, 1978. ISBN 0-201-02931-6 pbk.

None

Dr. J. C. Bezdek, Professor

Dr. Gautam Biswas, Assistant Professor

Dr. Terry Huntsberger, Assistant Professor

To introduce students to the basic techniques of Syntactic Pattern Recognition. Specifically, this course discusses the integration of numerical pattern recognition algorithms into higher level systems that perform clustering and classification using structural and semantic descriptors as well as facts and knowledge about the problem domain.

1. Numerical Pattern Recognition

2. Automata Theory

1.      Introduction, Overview, and History (1) 
2.      Formal Languages
        a.   Chomsky forms of Grammar (4)
        a.   Automata as Classifiers (4)
3.      String Languages
        a.      Pattern Primitives and Attributes (2)
        b.      PDL Grammars (2)
        c.      Character, Contour, Finger Print Grammars (3)
4.      Syntax Analysis
        a.      Parsing: Top Down vs. Bottom up (4)
        b.      CYK and Earley Parsers (4)
5.      Error Correcting Parsers for String Languages
        a.      MDECP for CF Languages (3)
        b.      MD Classifier Design (2)
        c.      Chromosome Grammars (1)
        d.      Bayes Classifiers in Syntactic Systems (2)
6.      Cluster Analysis
        a.      Nearest Neighbor/Nearest Prototype Methods (3)
        b.      MST Hiearchical method (1)
        c.      Syntactic distance measures (1)
7.      Applications in Image Processing, Shape Analysis, and Medicine (5)

Students will implement at least one syntactic pattern recognition algorithm, and use it to analyze a data set. Emphasis will be on derivation of primitives, attributes, parsing, and recognition. Typically, students will code and test, e.g., the CYK parser, nearest neighbor/nearest prototype clustering; or a syntactic Bayes classifier scheme.

1. Homework is assigned, collected and graded on a regular basis.
2. Two exams: one midterm, one final
3. Quizzes: intermittently

Estimated CSAB Category Content:

Not Applicable.

Syllabus Flexibility:

High

Top of the page