Identification High Influential Articles by Considering the Topic Characteristics of Articles

Abstract

The topic of one article reflects its main semantic content, which is also the main guidance for researchers to choose reference literature. In order to explore whether the topic of an article will affect its citation trend in future, this paper establishes a machine learning framework to study the role of topic characteristics in the prediction of future high influential articles. Articles from four different disciplines are collected as experimental samples to verify whether the framework proposed in this paper can be applied to the prediction task in different disciplines. The Latent Dirichlet Allocation (LDA) is used to determine the topic characteristics of sample articles. LDA can map sample articles to current hot topics and generate the mapping probability of sample articles under different hot topics. The maximum mapping probability of the sample article under the hot topics is extracted as the topic feature of the article. Then the feature space for the prediction task is constructed by combining the topic feature and some bibliometrics indices of articles. Three feature selection algorithms, Fisher Score, Relief-F and Spectral Feature Selection (SPEC), are taken to select the important features in the feature space. The prediction performance of these features is finally tested by three classifiers, SVM, KNN and Bagging. The experimental results show that the topic characteristics of article, the early citation characteristics of article, and the reputation of the author are the key factors that determine whether an article can grow into a highly influential one. The important value of topic characteristics in articles' citation activities shows that the content of the article is an important factor in attracting more citations.

Factors associating with or predicting more cited or higher quality journal articles: An Annual Review of Information Science and Technology (ARIST) paper

“…One or a range to compare, including linear regression (Saeed et al, 2008;Yu et al, 2014), and logistic regression (Fu & Aliferis, 2010;Ib añez et al, 2009), as well as classical machine learning algorithms like Support Vector Machines (SVM) (Fu & Aliferis, 2010;Wang et al, 2020), random forest (Robson & Mousquès, 2016), naïve Bayes (Ib añez et al, 2009), neural networks (Abrishami & Aliakbary, 2019;Wang et al, 2020), and deep learning designs (Ma et al, 2021;Xu et al, 2019). Statistical algorithms that rely on identify linear relationships tend to be less powerful, because less flexible, than most machine learning algorithms.…”

“…A third of a year (Falagas et al, 2013) or a single year (Wang et al, 2020; Yu et al, 2014) to 11 years (Chen & Zhang, 2015). Narrower ranges of years generate more powerful predictions due to increased homogeneity, especially if any of the data is not year normalized.…”

“…Early citations (Chen & Zhang, 2015; Ma et al, 2021; Ruan et al, 2020; Wang et al, 2020; Yu et al, 2014), citation graph (Cummings & Nassar, 2020; Zhao & Feng, 2022), number of authors (Lee, 2020; Yu et al, 2014), first/all author productivity (Lee, 2020; Yu et al, 2014), first/all author collaboration rates (Lee, 2020), gender (Haslam et al, 2008), first author country income level (Sin, 2011), other basic first/all author properties (Chen & Zhang, 2015; Fu & Aliferis, 2010; Haslam et al, 2008; Lokker et al, 2008; Wang et al, 2020; Yu et al, 2014), complex first/all author capabilities inferred from a matrix analysis of a large document set (Chen & Zhang, 2015; Lee, 2020), field citation rates (Chen & Zhang, 2015), reference count (Ha, 2022; Wang et al, 2020; Yu et al, 2014), reference impact (Boyack & Klavans, 2005), institution properties (Fu & Aliferis, 2010), abstract readability, abstract terms (Fu & Aliferis, 2010; Ibáñez et al, 2009), title terms (Fu & Aliferis, 2010), keywords (Fu & Aliferis, 2010), topics (Robson & Mousquès, 2016), title/abstract sentence semantic representations (Ma et al, 2021), study design (Falagas et al, 2013), journal section (Ibáñez et al, 2009), document type (BinMakhashen & Al‐Jamimi, 2022; Ha, 2022), altmetrics (Akella et al, 2021), journal self‐citation rate (Ruan et al, 2020), journal impact (Yu et al, 2014), other journal properties (Fu & Aliferis, 2010; Yu et al, 2014), language (Wang et al, 2020), page count (Robson & Mousquès, 2016), title length (Robson & Mousquès, 2016), abstract length (Lokker et al, 2008; Robson & Mousquès,…”

“…Yes (Wang et al, 2020) or No (Saeed et al, 2008). Feature selection refers to a procedure to select the most useful features from an initial set.…”

The Impact of Popular Science Articles by Physicians on Their Performance on Online Medical Platforms

“…The topic is a key factor in understanding an article’s content, and the probability of distribution of topics can be identified by machine learning techniques [ 24 ]. In the e-commerce field, by analyzing the thematic characteristics of the review content, we can summarize the importance that consumers attach to different features of the product, thereby helping merchants attract more potential consumers [ 25 ].…”