An Improved PageRank Algorithm Based on Text Similarity Approach for Critical Standards Identification in Complex Standard Citation Networks

Abstract

A standard system, which is a powerful tool in maintaining the normal operations and development of a specific industry, is intrinsically a complex network composed of numerous standards which coordinate and interact with each other. In a networked standard system, the identification of critical standards is of great significance when drafting and revising standards. However, a majority of the existing literature has focused on the citation relationships between standards while ignoring the intrinsic interdependent relationships between the contents of standards. To overcome this limitation, we utilize the text similarity approach (TSA) to quantify the relationship intensity between each pair of standards, in order to generate a directed weighted network. The critical contribution of this study is that the similarity computed by the TSA is incorporated into the traditional PageRank algorithm for the identification of critical standards. The improved algorithm comprehensively considers the quantity and importance of neighboring standards and the citation intensity, as quantified by TSA. The algorithm is finally validated using the Chinese environmental health standards through comparison with the traditional PageRank algorithm and different classic measurements.

A Scientific Paper Recommendation Framework Based on Multi-Topic Communities and Modified PageRank

“…Considering the citation functions, some researchers developed personalized PageRank rank papers for recommendation by modifying the bias probability vector and the transition probabilities matrix. Wei et al [19] and Xie et al [45] computed the relationship strength between each pair of nodes by utilizing a text similarity approach to generate a directed and weighted network and then incorporate it into the PageRank algorithm. Dunaiski et al [51] observed that personalizing PageRank with citation counts of papers decreases time bias but increases topic bias.…”

“…Unlike the original PageRank, the random walker of personalized PageRank jumps to specific nodes with certain weights driven by a bias probability vector and transition probability matrix [13]. Some researchers employed personalized PageRank to modify a bias probability vector and transition probability matrix, which corresponds to texts [19], time [20] [21], semantic [22], topic [23][24] [25], and time + topic [26] [27]. PageRank incorporates time and topic variables to deal with biased age and field [27].…”

The use of deep learning integrating image recognition in language analysis technology in secondary school education

“…In today’s teaching practice, there are often phenomena, such as reading from books and leaving textbooks, which seriously affect the improvement of teaching quality. The degree to which the teachers read or mechanically copy the textbook or courseware is defined as content similarity 25 . Further mining learners’ evaluation comments implies that most learners strongly oppose the high content-similar teaching behaviors, such as reading books or reading courseware.…”

Fusion of the word2vec word embedding model and cluster analysis for the communication of music intangible cultural heritage

“…Both texts are vectorized using the proper technology, and then the degree of similarity between them is determined using an appropriate algorithm. The similarity between two texts is shown by the magnitude of the numerical number returned by the calculation 5 . Text similarity algorithms are at the heart of numerous applications, including the categorization and grouping of texts, AI-powered question answering systems, and search engines.…”