Extracting Topics from Texts Based on Situations

217

Authors

nan

Abstract

To understand text, we must relate it with specified situations. This paper, on the basis of such an idea, discusses how the things that a text describes and the situation that the text relates to are expressed in a computer and how the topic of a text is extracted. 1.

Keywords

nan

Full Content

Language, Information and Computation(PACLIC 11), 1996, 357-362 Extracting Topics from Texts Based on Situations Ma Zhiyi,Zhan Xuegong,Yao Tianshun Department of Computer Science, Northeastern University Shenyang, 110006 China Abstract To understand text, we must relate it with specified situations. This paper, on the basis of such an idea, discusses how the things that a text describes and the situation that the text relates to are expressed in a computer and how the topic of a text is extracted. 1.Introduction The topic of a piece of writing is the central idea of it. People not only care for what it expresses, but also care for the most important idea of it. Therefore, if each piece of writing has a marked topic, then retrieval rate and retrieval precision of writings can be increased to a great extent. In computers, a piece of writing is a text. Only when a text relates to certain situations, can its topic be rightly extracted. On the basis of situations matching, the paper presents the method of extracting topics from texts. A situation is something's world state in a specified moment and environment. So, when a thing's attributes are described, the thing's situations must be referred to, and at every moment, reasoning only can be executed in a situation. This is why situations are built. A situation is a function of time and space. For example, things and their behaviors can constitute different situations at the same time and in difference space, or in the same space and at difference time by increasing or decreasing a little information. The situation knowledge of something is the sum of perceptual experience and rational knowledge about a topic. The degree of people understanding a text often relies on their knowledge level of the field to which the text refers. That is, same or similar situation knowledge is the foundation of people's understanding each other. Situations may be divided into broad sense situations and narrow sense situations. A broad sense situation includes all dynamic state events and static state events in the objective world. It is consecutive tableaus in consecutive time and space. A narrow sense situation is dynamic state or static state events that relates to a specified discourse, that is, it includes the environments in which the discourse happens and the events to which the discourse relates. The task of situation theory is extracting common inner structure from large amounts of real situations in the objective world, probing into the restrained relations among situations, revealing meaning of the language expressions, and presenting a calculable mathematical model for natural language understanding based on situations(Sun Bo 1992). A real situation is a set of specific dynamic or static events. An abstract situation is a set that isconstructed by some elements 357 with common inner structure extracted from large amounts of real situations, and it is a mathematical abstraction. Objects, relations among objects, time and space are all basic elements that establish an abstract situation, and objects and relations among objects are of …

Content truncated to 500 words

Labeled Text

To understand text, we must relate it with specified situations. This paper, on the basis of such an idea, discusses how the things that a text describes and the situation that the text relates to are expressed in a computer and how the topic of a text is extracted. 1.

Publication Details

Paper ID:
217
Year:
1996
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan