Intro
The FileBasedSpellChecker uses a flat file to build a Lucene contrib/spellcheck index. There is currently no frequency information used in calculating spelling suggestions, so the SpellCheckComponent will not be able to provide "only more popular" or other frequency based information. As a workaround, it isn't all that hard to create an index from a file and have it weight the terms.
Format
The file format is one word per line, as in
pizza history junk