Modified Key-Sentence Extraction by RICOH at NTCIR-2 TSC
نویسنده
چکیده
We participated at NTCIR-2 in the TSC subtask A1 and A-2 using the output of QJR/KSE, a function of key-sentence extraction. Through examining the evaluation results and the human-extracted sentence data of the dryrun subtask A-1, we tried to make an experimental hybrid system incorporating the lead method into the original. In the formal run, the evaluation results of the hybrid system in the subtask A-1 and A-2 were in the top class among the twelve systems, much better than the results obtained in the dryrun. Especially in the F-measures of the 30% summarization rate, our submitted two systems got the top and second rankings. As a result of our participating in TSC, we have examined some internal scores to study the effectiveness of their results and had a chance to make some improvements and tuned up the hybrid system to increase its performance.
منابع مشابه
TISS: An Integrated Summarization System for TSC-3
In consideration of the previous workshop, we participate in TSC-3 to make improvements on important sentence extraction used in dry run of TSC-2. We formulate important sentence extraction as a combinational optimization problem that determines a set of sentences containing as many important information fragments as possible. In addition to the extraction method, we reinforce peripheral compon...
متن کاملNTT/NAIST's Text Summarization Systems for TSC-2
In this paper, we describe the following two approaches to summarization: (1) only sentence extraction, (2) sentence extraction + bunsetsu elimination. For both approaches, we use the machine learning algorithm called Support Vector Machines. We participated in both Task-A (single-document summarization task) and Task-B (multi-document summarization task) of TSC-2.
متن کاملComparison of Feature Usage at TSC-3 Summarization Tasks
We participated in two summarization tasks at the TSC-3. We have introduced categorization of feature values for our summarization system, which is based on sentence extraction technique. The categorized values were used as features for generating a decision tree. We compared our summarization system using the categorization of feature values with the one using linear combination of features in...
متن کاملSentence Extraction System Assembling Multiple Evidence
We have developed a sentence extraction system that estimates the significance of sentences by integrating four scoring functions that use as evidence sentence location, sentence length, TF/IDF values of words, and similarity to the title. Similarity to a given query is also added to the system in the summarization task for information retrieval. Parameters for scoring functions were optimized ...
متن کاملRICOH at NTCIR-2
At NTCIR-2, RICOH submitted eight runs for the Japanese IR task. Of the eight runs, four runs use the title eld only and the other four use the description eld only. RICOH's system is built on our English text retrieval system and augmented to handle Japanese text. The system features (1) hybrid retrieval using a combination of n-gram indexing and wordbased document ranking; (2) word-based and ...
متن کامل