Headline Preprocessing

News headline text is scanned preliminarily to break it apart into text fragments of legitimate characters; each text fragment is a candidate for use as an indexed keyword. The following rules apply to the preprocessing scan:

Characters ‘A’ to ‘Z’, ‘a’ to ‘z’, and ‘0' to ‘9’ will be accepted as keyword components.

The period (‘.’), ampersand (‘&’), apostrophe (‘'’), and hyphen (‘-’) characters will be accepted if they do not begin the text fragment. (e.g., the text fragments “U.S.” or “T-Bonds” are acceptable as shown.)

All other characters will be treated as empty space between text fragments.