Abstract—The development of an application that can help the thesis adviser and student construct and recommend thesis titles through the use of Natural Language Processing (NLP) was the primary objective of this study. Resembling the concept of Information Retrieval under Query-Focused Summarization, words except stop words from the related literature and studies of the thesis and its title were used as a training corpus while the words except stop words of chapters one (1), two (2) and three (3) of the thesis or merely the test document acts as the query in the corpus, where it retrieved its weight from both training and test records. The result showed that out of forty-five (45) thesis titles, twenty-eight (28) different title formats were constructed by the thesis advisers with the support of the developed application. The advanced application obtained the accuracy score of 1.55 or “Accurate” and quality score of 4.06 or “Good” in fifteen (15) thesis titles.

 

Index Terms—thesis title, electronic resources, natural language processing, information retrieval, query-focused summarization

Cite: Frances Frangelico S. Friginal, Gerald T. Cayabyab, and Bartolome T. Tanguilig III, "Dynamic Weight-Based Approach for Thesis Title Recommendation," International Journal of Signal Processing Systems, Vol. 4, No. 6, pp. 528-536, December 2016. doi: 10.18178/ijsps.4.6.528-536