Title of Publication: A REVIEW OF CLASSIFICATION MODELS ARISING FROM FULL MULTINOMIAL RULE
Author(s): G.A. OSUJI, S.I. ONYEAGU
Year of Publication: 2009
In this paper, models arising from the Full multinomial rule were reviewed. The models are: The M-rule, the D-rule, the L-rule and the P-rule. Under different population structures and assuming equal cost of misclassification and a priori probabilities, we discovered that the rules still fall back to the Full multinomial rule. Although the problem of zero frequencies and state proliferations were taken care of in the rules; yet the Full multinomial rule is the most favouredKeywords:
Multinomial rules; State proliferation; Classification rules; Discrete random variables.
Need complete download? Login to download