International Journal of Computer Processing of Languages

Analyzing the logical structure of a sentence is important for understanding natural language. In this paper, we present a task of Recognition of Requisite Part and Effectuation Part in Law Sentences, or RRE task for short, which is studied in research on Legal Engineering. The goal of this task is to recognize the structure of a law sentence. We investigate the RRE task regarding both the linguistic features and problem modeling aspects. We also propose solutions and present experimental results in a Japanese legal text domain. We got 88.58% with a supervised learning model and 88.84% with a semi-supervised learning model in the F_β=1 score on the Japanese National Pension Law corpus.

The "love code" is often used by netizens to exchange short text messages over mobile devices, whose characteristic is one-character-one-key. In this paper, using 1986 collected love codes, the corresponding numerical digits (0~9) for 523 Chinese phonetic symbols are extracted first. Using the Chinese phonetic symbols in each numerical key and proposed similarity measure, the corresponding numerical digits for undecided Chinese phonetic symbols are determined. Meanwhile, 2065 Chinese canned short messages (approximate 1100,000 Chinese characters) are collected from the Internet and used as the training corpus. The n-gram language model is employed to develop the statistical Chinese short message input system. Finally, the experiments show promising results.

We introduce novel paraphrasing rules for Japanese light verb constructions (LVCs) that reduce the differences in the surface forms while retaining several of the crucial syntactic/semantic functions of these light verbs. An analysis of the linguistic properties of light verbs allows us to create paraphrasing patterns that map 151 different light verbs into 10 simple forms. Of these 10 forms, 7 convert complex noun-particle-verb structures into simple predicative forms. By constructing a list of 923 examples for ambiguous light verbs, we show that we can correctly distinguish real LVCs from those in which the light verbs were actually functioning as a main verb. The results of experiments indicate that our paraphrasing rules offer high accuracy. The experiments also reveal that our paraphrasing system works as a normalizer of complex predicates, which improves the recall rate of the predicate extraction task.

Due to the rapid change of the real world, it is impossible for a Web map service to gather all recently appeared addresses in its spatial database, so as to provide an effective on-demand service. This paper proposes a novel method that takes advantage of Web mining to locate unregistered toponyms by utilizing address information available on the Internet. The design principle of this method consists of two key steps. First, multiple Web pages that contain the unregistered toponym are collected, and the recognized addresses, which already exist in the spatial database, are extracted from the Web pages. Second, the spatial relationship between each recognized address and the unregistered toponym is calculated to determine which recognized addresses have the highest weights, and accordingly locate the unregistered toponym on the map with high accuracy.

In this paper, we propose a graph-based dependency parsing model using dynamic features. The dynamic features, which are derived from the partial tree during bottom-up chart parsing, play an important role in finding the correct heads. Based on the beneficial aspects of the head-final property of Korean, we suggest a variant of the CYK parsing algorithm with an O(n³) complexity which has the ability to search for the maximum spanning tree (MST) from all projective trees. Compared to Eisner's algorithm, this algorithm is time-efficient in parsing all other head-final languages. We also explain why the graph-based approaches are seldom attempted in parsing morphologically rich languages (MRLs) and suggest a morpheme-level dependency representation as a solution. Finally, we evaluate the proposed method on the KAIST corpus using two discriminative online learning methods (the averaged perceptron algorithm and passive-aggressive algorithm) and show that the trained parser achieves the best reported parsing accuracy.

In this paper, we focus on the reliability of information encoded in a Web 2.0 community platform. Specifically, we aim to explore how linguistically encoded clues can contribute to the task. Given that evidentiality is the linguistic representation for the reliability of a statement, we propose to use evidential, the lexicalized evidentiality, to model text representation under the framework of machine learning based text classification. Based on the model, we conducted experiments to identify the best answers in Collaborative Question Answering (CQA) services. The experimental results confirm that, incorporating evidential for predicting text reliability is effective, since it shows the writer's self-judgment for information reliability. Moreover, our method can largely reduce the dimensionality of the vector space, and therefore provide an improvement in efficiency.