Page 336 -
P. 336

310   Chapter 8 ■ Classification


                             Any reasonable scheme for merging the results from multiple classifiers
                           must deal with three important issues:
                             1. The response of the multiple classifier must be the best one given
                                the results of the individual classifiers. It should in some logical way
                                represent the most likely true classification, even when presented with
                                contradictory individual classifications.
                             2. The classifiers in the system may produce different types of response.
                                These must be merged into a coherent single response.
                             3. The multiple classifier must yield the correct result more often than any
                                of the individual classifiers, or there is no point.
                             The first problem has various potential solutions for each possible type of
                           response, and these will be dealt with first.

                           8.5.2    Merging Type 1 Responses
                           Given that the output of each classifier is a single, simple classification value,
                           the obvious way to combine them is by using a voting strategy. A majority
                           voting scheme can be expressed as follows: let C i (x)be the result produced by
                           classifier i for the digit image x, where there are k different classifiers in the
                           system; then let H(d) be the number of classifiers giving a classification of d for
                           the digit image x,where d is one of {0,1,2,3,4,5,6,7,8,9}. H can be thought of as
                           a histogram, and could be calculated in the following manner:

                             for (i=0; i<k; i++)
                                    H[ Ci(x) ] += 1;
                             Then, the overall classification E, expressing the opinions of the k classifiers,
                           could be:
                                            
                                                                              k
                                              j   if max(H(i)) = H(j)and H(j) >
                                            
                                     E(x) =                                   2            (EQ 8.8)
                                              10            otherwise
                                            
                             This is called a simple majority vote (SMV). For comparison, a parliamentary
                           majority vote would simply select j so that H(j) was a maximum. An easy
                           generalization of this scheme replaces the constant k/2 in the above expression
                           with k*α for 0 <= α<= 1 [Xu, 1992]. This permits a degree of flexibility
                           in deciding what degree of majority will be sufficient, and will be called a
                           weighted majority vote (WMV). This scheme can be expressed as:
                                            *
                                             j   if max(H(i)) = H(j)and H(j) >αk
                                     E(x) =                                                (EQ 8.9)
                                             10             Otherwise
                             For example, many important votes in government and administrative
                           committees require a 2/3 majority in order to pass. This would be equivalent
                           to a value of α = 2/3inEquation8.9.
   331   332   333   334   335   336   337   338   339   340   341