Abstract:
The paper suggests a view on categorizing text expressions that are generally referred to by information extraction community as time-point expressions, or temporal coordinates. Two types of expressions are identified which differ in the way they refer to time.
The issues of normalization (i. e. identifying the absolute value) are addressed for both types of expressions. (In Russian).
Key words and phrases:natural language processing, temporal information extraction, normalization of context-dependent time expressions.