Corpus Database

The corpus database consists of the corpus system and corpus data

The corpus system software was developed by a research group from the Computer-Aided Translation Unit of Universiti Sains Malaysia under the USM-DBP collaboration framework based on a memorandum of understanding signed in 1993.

This corpus system is equipped with the feature to access word forms, derivatives and phrases. The search results are displayed in the form of concordance lines with the search keywords organised and sorted in the middle of the line.
 
A search based on these keywords can be made through a variety of search techniques according to the information to be extracted and displayed. The common search techniques used are as follows:

a) Access via Keyword (Word Form)

A word form can be accessed by typing the word. For example, a search that uses the keyword “kata” will display all forms of this word that exist in a text corpus. (See screen display example)

b) Access via Keyword and the symbols “*” and “?”

Accessing a word form can also be done by using a keyword and the symbols “*” and “?” (“*” represents one or more characters and “?” represents one character).

For example, a search using the keyword “*kata*” will display forms such as “kata,” “perkataan,” “berkata” and so on. (See screen display example)

A search using the keyword “b?t?l” will display forms such as “botol,” “batal,” “betul” and so on. (See screen display example)

Text analysis
The corpus system is also equipped with a module to analyse texts (known as MATA - Malay Text Analysis) and generate statistics on a text as follows:

(a)  Number of words;

(b)  Word frequency;
(c)  Number and list of root words;

(d)  Number and list of new words; and
(e)  Number and list of non-authentic words.

Corpus Data

DEFINITION

In general, corpus can be defined as “a collection of articles (writings, etc.) regarding a particular matter or collection of materials to be studied (such as group of word usage, etc.)” (Kamus Dewan, third edition, 1994). However, in modern linguistics, “corpus” has the additional meaning of material that is “read and processed by a computer”.

This means that the corpus collected and maintained in this project is a collection of digital texts that can be processed using the techniques and methods of linguistic computing to show patterns and to correlate one word with another.

DATA

Corpus data may be derived from written or oral form. However, the current focus of this programme is still on written material from books, magazines, newspapers, monographs, documents, working papers, correspondence, brochures, and so on.

Each type of discourse is compiled in a separate sub-corpus. As of 25 November 2008, the Corpus Database had approximately 135 million words contained in the following ten sub-corpuses:

CORPUS-BASED GOALS AND RESEARCH OUTCOMES

This database was built with the aim of providing research data that can be used for compiling dictionaries, grammar research and other forms of linguistic studies. Examples of corpus-based studies are as follows: