Bayesian estimation is known to improve recognition accuracy.
However, the method is not effective due to the bias of
the probability when sample sizes are not the same.
In this paper, the function of the bias of Geisser's predictive
distribution is derived without any approximation.
It is shown that correcting likelihood by the function improves
recognition accuracy.
This indicates the bias is one of the cause of decreasing recognition
accuracy.
From the character of the function, it is easy to explain the
knowledges obtained from experiments.
The derivation of the function can be applied to other predictive
distributions.
Tab-separated. Paste into the sheet starting at column A. Required fields that cannot be inferred are left blank.
Format definition
Edits apply immediately and stay in effect until you pick a style again or reset. First line is the default; add lines like "book: …" to override per type.
How to write the definition
Insert a field with braces like {authors}. A part enclosed in [ ... ] is omitted entirely when any field inside it is empty. Available fields:
One macro per line as \name{field}{field}…. A field name in braces is replaced by its (LaTeX-escaped) value; other text is kept. Rename macros or add lines freely — they appear in the macro selector. A line like master_thesis: thesis fixes the macro for a type (automatic choice: journals use paperdoi with a DOI, paperno with a number, else paper; conferences conf with pages, else confpaper; books book). {pages} comes as 119--124; {journal} falls back to the book title. Fields: