SENTENCE ANALYSIS WITH ARTIFICIAL INTELLIGENCE MACHINE LEARNING USING FINITE STATE AUTOMATA

Abstract

Sentence pattern in Indonesian is very important. Incorrect pattern can cause
misinterpretation and different meaning. The basic sentence pattern consists of subject,
predicate, and object (SPO). To analyze the sentence, this project implements Finite State
Automata (FSA) algorithm, which is useful to identify the class of words in a sentence.
The given sentence processed through parsing process to break down the sentence into
array of tokens. The program will then examine the tokens to determine the class of each
word. The discovered pattern will be checked against existing vocabularies on the
database. If the word exists, the program will return a response in accordance to the
existing one. Otherwise, the program will store the word onto a temporary file until it find
the class of the word.
The Intelligence of the program depends on the number of vocabularies on the database.
This program supports learning process, so the more sentences given, the more intelligent
it gets.