(3 intermediate revisions by the same user not shown)
Line 4: Line 4:
  
  
In the second lecture, we
+
In the second lecture, we discussed the difference between finite and infinite feature spaces. We saw that, when the feature space of a decision problem is finite, then one does not necessarily need to use the decision theory techniques we will covered in [[ECE662]]. More specifically, if the feature space is relatively small, then one may be able to pre-compute the decisions for all possible feature values, and store the results in a look-up-table.
  
 +
In the case of infinite feature spaces, we discuss the use of [[Lecture_2_-_Decision_Hypersurfaces_OldKiwi|hypersurfaces]] in the feature space as a means of defining the separation between classes.
  
  
 +
Previous: [[Lecture1ECE662S10|Lecture 1]]
 +
Next: [[Lecture3ECE662S10|Lecture 3]]
 
----
 
----
 
[[OutlineECE662S10|Back to course outline]]
 
[[OutlineECE662S10|Back to course outline]]

Latest revision as of 08:03, 12 April 2010


Details of Lecture 2, ECE662 Spring 2010

In the second lecture, we discussed the difference between finite and infinite feature spaces. We saw that, when the feature space of a decision problem is finite, then one does not necessarily need to use the decision theory techniques we will covered in ECE662. More specifically, if the feature space is relatively small, then one may be able to pre-compute the decisions for all possible feature values, and store the results in a look-up-table.

In the case of infinite feature spaces, we discuss the use of hypersurfaces in the feature space as a means of defining the separation between classes.


Previous: Lecture 1 Next: Lecture 3


Back to course outline

Back to 2010 Spring ECE 662 mboutin

Back to ECE662

Alumni Liaison

has a message for current ECE438 students.

Sean Hu, ECE PhD 2009