Includes bibliographical references.

Reflections on the Awareness and Progress of Natural Language Processing (NLP) Research in the Philippines -- Clustering Different Language using Trigrams -- Towards an Extended Hybrid N-gram Grammar Checker Algorithm Applied in Filipino -- Automatic Text Summarization using ChainRank: A Hybrid Technique -- Using Statistical Machine Translation for Style ang Grammar Correction -- Meron o Mayroon: An Exploratory Study on Hiligaynon and Tagalog "Dipthongs"

[Article Title: Reflections on the Awareness and Progress of Natural Language Processing (NLP) Research in the Philippines / Rodolfo C. Raga, p.1-9]

Abstract: Natural Language Processing (NLP) is an area of research and application that explores how computers can be used to understand and manipulate natural language text or speech to do useful things. The number of top tech companies pouring funds into NLP research as well as the growing number of reputable foreign universities offering NLP courses makes it evident that NLP plays a crucial role in the business arena today. At the same time, if harnessed and properly directed, NLP technologies can be used as an instrument in addressing the language problems that afflict most multilingual countries. However, despite these potentials, in the Philippines, it can be observed that the level of awareness of students in this field of research seemingly remains low. A low level of awareness can potentially have inhibiting effects on the progress of this field of research. This article attemse in provide a thumbnail sketch of this situation.

[Article Title: Clustering Different Language using Trigrams / Angelica Dela Cruz, Maria Cristina Co, Adrian Martin Sy, and Nathaniel Oco, p.10-17]

Abstract: Proper discernment between closely related languages is ons problem in language identification. Strong similarities between certain languages could result in lower recall rates. In this paper, we clustered 43 Philippine languages by using trigrams to identify closely-related languages, and constructed a language family tree that conveys the origin of the languages. We collected online religious text documents and used 100,000 words for the 42 domain languages as training data. For the Yami language, a dictionary project, short story and a few transcribed verses were used. These were cleaned and character trigrams were generated.
The languages were clustered using two algorithms, farthest first algorithm and simple k-means algorithm.

[Article Title: Towards an Extended Hybrid N-gram Grammar Checker Algorithm Applied in Filipino / Matthew Philip GO and Allan Bora, p.18-23]

Abstract: In this research, we discuss an extension to an existing hybrid n-gram grammar checking algorithm (Tsao & Wible, 2009) to increase the coverage of the grammatical errors that such approach can solve. This extended approach is applied in the Filipino language where specific error types are present. The existing algorithm uses words, lemmas, and part-of-speech (pos) tags to create rules and correct grammar errors and suggest possible correction. It is also limited to just errors that can be corrected by replacing a word with another word.

[Article Title: Automatic Text Summarization using ChainRank: A Hybrid Technique / Michael Angeles, Nathaniel Laxina, Maria Art Antonette Clarino, and Riza Mercado, p.24-30]

Abstract: This paper presents an automatic text summarizer for English articles that uses ChainRank - a hybrid approach to text summarization based on combining Lexical Chains and PageRank. Given a text document, the system extracts the essential sentences that capture the whole idea of the document according to an evaluation process where sentences are scored.

[Article Title: Using Statistical Machine Translation for Style ang Grammar Correction / Manolito Octaviano, Jr., NAthaniel Oco, and Allan Bora, p.31-36]

Abstract: This paper presents the use of SMT approach for style and grammar checking. The principle behind the approach is to surface the common mistakes made by native speakers. Translation documents of English to Filipino were collected containing errors that were marked and corrected.

[Article Title: Meron o Mayroon: An Exploratory Study on Hiligaynon and Tagalog "Dipthongs" / Diane Manzano and Samantha jade Sadural, p.37-45]

Abstract: he vocoid sequences /aj, /uj/, /aw/, and /iw/ are observed to l widely present in most Philippine-type languages and considered by most historical linguists, like Blust and Ross, diphthongs. They, however, behave rather differently from th common diphthongs as they do not behave as a unit but as tw separate entities. The paper sides with Clynes' proposition an argues against the analysis of these vocoid sequences diphthongs. Departing from his study, the phonetic an phonological nature of these sequences are explored by plottin and analyzing the changes in formant frequencies of data fror two specific Philippine-type languages, Tagalog and Hiligaynon.