Prefix Extraction of Malay Words using Backpropagation Neural Network

This paper presents a different approach in extracting prefixes from Malay words, by training a backpropagation neural network to learn a set of rules. Since Malay prefixes change a rootword whereas suffixes don't, our attempt focuses on extracting prefixes from words. A set of rules are devised, on which the network is trained to a satisfactory level. Next, the network is tested upon using some samples. Then, an actual Malay text is fed into the network (one word at a time) to see how well the prefix-extraction performs. Results are analyzed and discussed.