Basic architecture, methods and algorithms of system for temporal information extraction from natural language texts

The given paper presents a basic architecture of the system which intended for temporal information extraction from natural language texts. We present the main structural parts of this architecture along with the corresponding methods and algorithms. Also we distinguish three stages in the temporal information extraction: recognizing the events and times, identifying the dependencies between them and final ordering of temporal elements. Correspondingly, logical inference, machine learning methods and linguistics banks of data are necessary in all these stages. (In Russian).