June-Jei Kuo

A summarization system for Chinese news from multiple sources

This article proposes a summarization system for multiple documents. It employs not only named entities and other signatures to cluster news from different sources, but also employs punctuation marks, linking elements, and topic chains to identify the meaningful units (MUs). Using nouns and verbs to identify the similar MUs, focusing and browsing models are applied to represent the summarization results. To reduce information loss during summarization, informative words in a document are introduced. For the evaluation, a question answering system (QA system) is proposed to substitute the human assessors. In large-scale experiments containing 140 questions to 17,877 documents, the results show that those models using informative words outperform pure heuristic voting-only strategy by news reporters. This model can be easily further applied to summarize multilingual news from multiple sources.

Cross document event clustering using knowledge mining from co-reference chains

Unification of the terminology usages which captures more term semantics is useful for event clustering. This paper proposes a metric of normalized chain edit distance to mine controlled vocabulary from cross-document co-reference chains incrementally. A novel threshold model that incorporates time decay function and spanning window utilizes the controlled vocabulary for event clustering on streaming news. The experimental results show that the proposed system has 16% performance increase compared to the baseline system and 6% performance increase compared to the system without introducing controlled vocabulary.

Cross-document event clustering using knowledge mining from co-reference chains

Unifying terminology usages which captures more term semantics is useful for event clustering. This paper proposes a metric of normalized chain edit distance to mine, incrementally, controlled vocabulary from cross-document coreference chains. Controlled vocabulary is employed to unify terms among different co-reference chains. A novel threshold model that incorporates both time decay function and spanning window uses the controlled vocabulary for event clustering on streaming news. Under correct co-reference chains, the proposed system has a 15.97% performance increase compared to the baseline system, and a 5.93% performance increase compared to the system without introducing controlled vocabulary. Furthermore, a Chinese co-reference resolution system with a chain filtering mechanism is used to experiment on the robustness of the proposed event clustering system. The clustering system using noisy co-reference chains still achieves a 10.55% performance increase compared to the baseline system. The above shows that our approach is promising.

A Library Recommender System Using Interest Change over Time and Matrix Clustering

A traditional library recommender system can not only employ individual history of library usage to recommend books which she (he) is interested, but also uses library usages of other users who are in the same social network to recommend the books which she (he) never loans but may be interested in. However, the same treatment for the user library usage at different times will lead to the recommended result departure from the users’ current information needs. Meanwhile, the data of library usage are highly dimensional and sparse. Thus, due to data sparsity and interest change over time, the traditional recommender systems cannot perform well. In order to tackle the two issues, this paper exploits time decay weight and matrix clustering to propose a novel library recommender system. Comparing two traditional recommender systems using K-Means clustering and hierarchical agglomerative clustering, experiments show promising results.

An Automatic Library Data Classification System Using Layer Structure and Voting Strategy

This paper deals with issues of traditional one-layered book classification systems and employs the complementary attribute of various classifiers to propose a two layered book classification system using voting strategy. Moreover, the collection of dissertations from a university library and books from an electronic bookstore are used as the training and testing corpus. The classification codes of dissertations and books are employed as the gold standard as well. Each dissertation contains various components such as title, authors, table of contents, abstract or cited papers et al. To understand the classification effect of all the combinations of components, various combinations are studied as well and the best combination is recommended. The features extracted from abstracts and table of content are found to be most useful for document classification. On the other hand, to obtain the best classification performance, the combination of classifiers for a two-layered book classification system is studied and the best combination was also recommended as well.

Question type classification and its application to a question answering system

-To explore the feasibility of semantic focus of inquiring information on a specific domain, this paper proposes a question classification algorithm using question word, intention word and some related words. An air traffic information service corpus was used to train a question type decision tree. An experiment showed that we were able to achieve 84% accuracy. Then we employed the question type classification algorithm to a question answering system for railway information service. Compared with other training methods like unigram or bigram, the accuracy could be improved from 65% to 70%. To deal with the domain shift problem, we adapted the question type classifier with a small railway corpus, the accuracy could be further improved to 80%. Besides the classification, we also disambiguate the semantics of some important temporal and spatial keywords. The experiments have shown that our method is promising.