A Directional Recognition Algorithm of Semantic Relation for Literature-Based Discovery
Literature-Based Discovery (LBD), a kind of knowledge discovery algorithm, is proposed by Don R. Swanson, which can assist the researchers to recognize implicit knowledge connection and further accelerate the generation of new knowledge. However, most of
- PDF / 311,651 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 0 Downloads / 178 Views
Abstract. Literature-Based Discovery (LBD), a kind of knowledge discovery algorithm, is proposed by Don R. Swanson, which can assist the researchers to recognize implicit knowledge connection and further accelerate the generation of new knowledge. However, most of algorithms in the field of LBD mainly start from the co-occurrence of terms to find connections between terms, and barely consider the semantic relation actually existing between pairs of terms. In this paper, a kind of directional recognition algorithm of semantic relation is put forward to recognize the directionality of semantic relation existing between pairs of terms. This algorithm will automatically judge the direction of semantic relation based on WordNet and JWNL. The numerical experiment results have indicated that the algorithm proposed in this paper can well recognize the directionality of the semantic relation. Keywords: Data mining WordNet JWNL
Natural language processing Semantic relation
1 Introduction Don R. Swanson was firstly proposed Literature-Based Discovery (LBD), a kind of knowledge discovery algorithm in 1986. Through development for about 30 years, many scholars have participated into the research on this algorithm, which has greatly improved and enhanced LBD. The algorithm can release the researchers from the limitations of the familiar and narrow research field, and we can avoid islanding phenomenon by virtue of this algorithm and efficiently support interdisciplinary intersecting innovation. The generalized discovery procedure of LBD proposed by Swanson can be briefly described as below. Find the intermediate word set B which co-occurs with A after starting from document set containing term A. Then further search document set b by starting from B, and recognize term set C which co-occurs with B from b. Later, starting from A and C to carry out the third search, if A and C occurs simultaneously in the same document, delete C, and if not, it can conclude there is recessive connection between A and C. And the obtained A-B-C is a kind of recessive connection knowledge. Based on research of Swanson, some scholars have successively raised some improved LBD algorithms. Yetisgen-Yildiz and Pratt have © Springer International Publishing Switzerland 2016 Y. Tan et al. (Eds.): ICSI 2016, Part II, LNCS 9713, pp. 281–288, 2016. DOI: 10.1007/978-3-319-41009-8_30
282
X. Liu et al.
proposed an open LBD system, - LitLinker [1]. The algorithm starts from the initial concept-Starting Term (e.g. migraine-), and finds the word that directly links with the initial concept, - linking terms. Then it will carry out information retrieval through linking terms and conducts another text mining on the obtained document to recognize each word that links with the linking term, - target terms. Weeber has developed a Literby system [2]. This algorithm realizes drug discovery by integrating the drug discovery process of Vos and LBD process of Seanson. This system is the extension of DAD system and still adopts the Two-Step knowledge discovery algorithm
Data Loading...
