Future Text Processing

-A look through Auto Indexing and Story Telling in the Domain of Pattern Recognition

ANWAR.N,
Department of Computer Science,
University of Calicut.

Many data mining techniques have been proposed for mining useful patterns
in text documents. How to effectively use and update discovered patterns is
still an open research issue, especially in the domain of text mining.

Auto Indexing:-

In the domain of text books the index is manually created in the written format
or in printed format. It is of the format pdf, txt,doc etc... In these, all the time
creating a document, we need to create the indexes also. Rather than that we
cannot go through the content pages from the index page. We need to
manually scroll down or up.
In this paper I propose an algorithm which takes document as input and
generate index files. In that index menu we can jump for a particular page or
particular segment of the document.

Story Telling:-

Recognized patterns may have many usages than the proposed one. We get
many key entities in going through the document. So a very small amount of
these key entities are used in indexing. The other values can be used for
creating a conclusion or summary of the document.
In another sense randomly selected words can be used for story telling.
Combining these words with some words from a predefined dictionary and
create meaningful sentences which form a story. So from a selected set of
words which can be arranged in many ways and create multiple stories. Also
a same story can be said in multiple ways.