Abstract—Detecting code-switching points is important, especially with the increasing globalism and multilingualism. However, this is a challenging task, but with the help of computers and technology, this can be done easily. In this paper, an approach to effectively detect code-switching points in a Tagalog-English text input, especially those with alternating English and Tagalog words, is presented. The approach uses the frequency counts of word bigrams and unigrams from language models which were trained from an existing and available corpus. For the testing, 3 test data categories were used – twitter posts, conversations, and short stories. The test data were composed of a total of 3088 English and Tagalog words. The results show that the system’s accuracy of properly identifying English and Tagalog words ranged from 81% - 95%, while the F-measure ranged from 72% - 95%. The research can be extended and improved using other n-grams, stemming, and searching algorithms.

Index Terms—Code-switching point detection, intra-sentential code-switching, word bigram, word unigram.

A. C. Bacatan, B. L. Castillo, M. J. Majan, and V. Palermo are with the University of Santo Tomas, Manila, Philippines (e-mail: ariannabacatan@gmail.com, bryan.loren.castillo@gmail.com, marjoriemajan@gmail.com, lian.palermo@gmail.com).
R. Sagum is with the Polytechnic University of the Philippines, Sta. Mesa, Manila and is also with the University of Santo Tomas, Manila (e-mail: riasagum31@yahoo.com).

Cite:Arianna Clarisse R. Bacatan, Bryan Loren D. Castillo, Marjorie Janelle T. Majan, Verlia F. Palermo, and Ria A. Sagum, "Detection of Intra-Sentential Code-Switching Points Using Word Bigram and Unigram Frequency Count," International Journal of Computer and Communication Engineering vol. 3, no. 3, pp. 184-188, 2014.