Predicting the citation counts of individual papers via a BP neural network

Predictable by publication: discovery of early highly cited academic papers based on their own features

Purpose

Predicting highly cited papers can enable an evaluation of the potential of papers and the early detection and determination of academic achievement value. However, most highly cited paper prediction studies consider early citation information, so predicting highly cited papers by publication is challenging. Therefore, the authors propose a method for predicting early highly cited papers based on their own features.

Design/methodology/approach

This research analyzed academic papers published in the Journal of the Association for Computing Machinery (ACM) from 2000 to 2013. Five types of features were extracted: paper features, journal features, author features, reference features and semantic features. Subsequently, the authors applied a deep neural network (DNN), support vector machine (SVM), decision tree (DT) and logistic regression (LGR), and they predicted highly cited papers 1–3 years after publication.

Findings

Experimental results showed that early highly cited academic papers are predictable when they are first published. The authors’ prediction models showed considerable performance. This study further confirmed that the features of references and authors play an important role in predicting early highly cited papers. In addition, the proportion of high-quality journal references has a more significant impact on prediction.

Originality/value

Based on the available information at the time of publication, this study proposed an effective early highly cited paper prediction model. This study facilitates the early discovery and realization of the value of scientific and technological achievements.

Measuring knowledge contribution performance of physicians in online health communities: A BP neural network approach

Extant literature on measuring the performance of physicians’ knowledge contribution in an online health community (OHC) is limited. To address this gap, this article aims to (1) develop a measurement model for physicians’ knowledge contribution performance; (2) use BP neural network to assign reasonable weight to each indicator of the model; and (3) explore the status and differences of knowledge contribution performance among a group of physicians. Based on the sample of 5407 infectious disease physicians in a Chinese OHC, we propose the measurement model by integrating physicians’ active knowledge contribution (AKC) and responsive knowledge contribution (RKC), covering 11 dimensions of contribution quantity and quality. We employ the BP neural network to optimise the model weights using the initial weight of the model obtained by the entropy method. The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method is used to evaluate the performance of physicians’ knowledge contribution in the OHC. The results show that it is feasible to use BP neural network to assign model weights. The distribution of physicians’ knowledge contribution performance is uneven; only a few have a high-level knowledge contribution performance. Meanwhile, a significant positive correlation exists between a physician’s title and respective knowledge contribution performance. Our research may contribute to related literature and practices by offering a fine-grained understanding of the performance of physicians’ knowledge contribution.

Students’ adaptive deep learning path and teaching strategy of contemporary ceramic art under the background of Internet +

With the rapid expansion of Internet technology, this research aims to explore the teaching strategies of ceramic art for contemporary students. Based on deep learning (DL), an automatic question answering (QA) system is established, new teaching strategies are analyzed, and the Internet is combined with the automatic QA system to help students solve problems encountered in the process of learning. Firstly, the related theories of DL and personalized learning are analyzed. Among DL-related theories, Back Propagation Neural Network (BPNN), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) are compared to implement a single model and a mixed model. Secondly, the collected student questions are selected and processed, and experimental parameters in different models are set for comparative experiments. Experiments reveal that the average accuracy and Mean Reciprocal Rank (MRR) of traditional retrieval methods can only reach about 0.5. In the basic neural network, the average accuracy of LSTM and GRU structural models is about 0.81, which can achieve better results. Finally, the accuracy of the hybrid model can reach about 0.82, and the accuracy and MRR of the Bidirectional Gated Recurrent Unit Network-Attention (BiGRU-Attention) model are 0.87 and 0.89, respectively, achieving the best results. The established DL model meets the requirements of the online automatic QA system, improves the teaching system, and helps students better understand and solve problems in the ceramic art courses.

A modified LSTM network to predict the citation counts of papers

Quantifiable predictability in the citation counts of articles is significant in scientometrics and informetrics. Many metrics based on the citation counts can evaluate the scientific impact of research articles and journals. Utilising time series models, an article’s citation counts up to the yth year after publication can be predicted by those up to the previous years. However, the typically used models cannot predict the fat tail of the actual citation distributions. Thus, based on cumulative advantage of the citation behaviour, we propose a method to predict the accumulated citation counts, by using a random number sampled from a power-law distribution to modify the results given by a recurrent neural network (RNN), long short-term memory. Extensive experiments on the data set including 17 journals in information science verified the effectiveness of our method by the good fittings on distributions and evolutionary trends of the citation counts of articles. Our method has the potential to be extended to predict other popular assessment measures such as impact factor and h-index for journals.

A Machine Learning Model to Predict Citation Counts of Scientific Papers in Otology Field

One of the most widely used measures of scientific impact is the number of citations. However, due to its heavy-tailed distribution, citations are fundamentally difficult to predict but can be improved. This study was aimed at investigating the factors and parts influencing the citation number of a scientific paper in the otology field. Therefore, this work proposes a new solution that utilizes machine learning and natural language processing to process English text and provides a paper citation as the predicted results. Different algorithms are implemented in this solution, such as linear regression, boosted decision tree, decision forest, and neural networks. The application of neural network regression revealed that papers' abstracts have more influence on the citation numbers of otological articles. This new solution has been developed in visual programming using Microsoft Azure machine learning at the back end and Programming Without Coding Technology at the front end. We recommend using machine learning models to improve the abstracts of research articles to get more citations.

COVID-19 citation pandemic within the psychological knowledge domain

Presented study explores the knowledge domain of psychological research published in 2020 and 2021. Metadata from 156,942 psychology papers available in Scopus were analyzed using citation analysis and bibliographic mapping techniques. Having in mind the ubiquity of the COVID-19 pandemic and the numerous ways it has affected people's lives, the fact that COVID-19-related papers represent only 2% to 7% of the total academic production in psychology may seem rather low. However, these papers have attracted much more attention from the public than non-COVID papers. They were also cited two to eight times more frequently, depending on the measure used, and account for 16% to 19% of total citations to psychology papers. Results show that early-stage researchers and those who had fewer articles in Scopus have benefited more from publishing COVID papers. They have managed to boost their average citation rates to the level of their colleagues who were much longer active and previously had higher citedness. Results indicate that the authors citing behavior largely follows the cumulative advantage pattern. Psychological research in general is mainly focused on mental health, anxiety, depression, and stress. This trend is even more fostered due to the pandemic since some of these topics are often analyzed as typical emotional reactions to COVID-19. Other relevant issues are also very well covered, except for the question how scientific results are communicated to the public. The role of "hot" papers was elaborated from the perspective of research evaluation practice.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12144-022-03146-3.

The relationship between open technological innovation, intellectual property rights capabilities, network strategy, and AI technology under the Internet of Things

The problems faced by the open technological innovation of China’s new ICT (information and communications technology) industry under IoT (Internet of Things) technology are expected to be analyzed to improve the overall innovation ability and ensure the sustainable development of related industries. An evaluation model is constructed for open technological innovation in the IoT industry by analyzing the development of IPR (Intellectual Property Rights) management, network strategy, and AI (artificial intelligence) technology under the development of IoT technology. Meanwhile, IBM SPSS Statistics 20.0 and IBM SPSS Amos 19.0 are used to analyze the data information of 306 enterprises in the information technology industry. Besides, the proposed hypotheses are verified by factor analysis, multiple regression, and Back Propagation Neural Network. Finally, a new evaluation index system is constructed for open technological innovation in the new ICT industry. The development of IoT technology provides a primary guarantee for the open technological innovation of the new ICT industry, and the network strategy has the greatest influence on the internal knowledge output mode. Besides, the experimental results indicate that the IoT and artificial intelligence have a critical display value for the open technological innovation of the emerging ICT industry, with the highest weight ratio of 48.25%. This result demonstrates that artificial intelligence is positively correlated with the external input information. Intellectual property management is a crucial guarantee of open technology innovation in the ICT industry. The evaluation model of open technological innovation in the ICT industry has good performance through case analysis, with the highest accuracy of 91.25%. Therefore, the evaluation index system reported here can reflect the important factors affecting the development of innovative technology, which can provide a theoretical basis and practical value for improving the existing open technology innovation system.

Performance Evaluation of Knowledge Sharing in an Industry-University-Research Alliance Based on PSO-BPNN

Knowledge sharing performance is very important to evaluate the interests of industry university research alliance. Firstly, this paper puts forward the index system of knowledge sharing performance evaluation of industry university research alliance. A BP neural network (BPNN) and a PSO-improved BP neural network (PSO-BPNN) are used to establish the evaluation model, and the accuracy of the model is evaluated. The prediction accuracy is evaluated by comparing the predicted value with the target value. Through comparative analysis, it is found that the performance evaluation model based on PSO-BP neural network has high accuracy and applicability, and is an effective method of alliance knowledge sharing performance evaluation.

Can the quality of published academic journal articles be assessed with machine learning?

Formal assessments of the quality of the research produced by departments and universities are now conducted by many countries to monitor achievements and allocate performance-related funding. These evaluations are hugely time consuming if conducted by post-publication peer review and are simplistic if based on citations or journal impact factors. This article investigates whether machine learning could help reduce the burden of peer review by using citations and metadata to learn how to score articles from a sample assessed by peer review. An experiment is used to underpin the discussion, attempting to predict journal citation thirds, as a proxy for article quality scores, for all Scopus narrow fields from 2014 to 2020. The results show that these proxy quality thirds can be predicted with above baseline accuracy in all 326 narrow fields, with Gradient Boosting Classifier, Random Forest Classifier, or Multinomial Naïve Bayes being the most accurate in nearly all cases. Nevertheless, the results partly leverage journal writing styles and topics, which are unwanted for some practical applications and cause substantial shifts in average scores between countries and between institutions within a country. There may be scope for predicting articles scores when the predictions have the highest probability.

Peer Review

https://publons.com/publon/10.1162/qss_a_00185

Is everyone invited to the discussion table? A bibliometric analysis COVID-19-related mental health literature

BACKGROUND

The COVID-19 pandemic has captured the mental health discussion worldwide. Examining countries' representation in this discussion could prove instrumental in identifying potential gaps in terms of ensuring a truly global conversation in times of global crisis.

METHODS

We collected mental health and COVID-19-related journal articles published in PubMed in 2020. We focused on the corresponding authors' countries of affiliation to explore countries' representation. We also examined these articles' academic impact and correlations with their corresponding authors' countries of affiliation. Additional journals and countries' indicators were collected from the Web of Science and World Bank websites, respectively. Data were analyzed using the IBM SPSS Statistics and the VOSviewer software.

RESULTS

In total, 3492 publications were analyzed. Based on the corresponding author, high-income countries produced 61.9% of these publications. Corresponding authors from Africa, Latin America and the Caribbean, and the Middle East combined accounted for 11.8% of the publications. Europe hosted corresponding authors with the most publications and citations, and corresponding authors from North America had the largest mean journal impact factor.

CONCLUSIONS

The global scientific discussion during the COVID-19 pandemic saw an increased contribution of academics from developing countries. However, authors from high-income countries have continued to shape this discussion. It is imperative to ensure the active participation of low- and middle-income countries in setting up the global mental health research agenda, particularly in situations of global crisis, such as the ongoing pandemic.

Innovation Strategy of Cultivating Innovative Enterprise Talents for Young Entrepreneurs Under Higher Education

A time series model is designed based on the backpropagation neural network to further optimize the innovation and development of new ventures. The specific situation of two factors is primarily analyzed as follows: the supply and demand ratio of enterprise talents and the state of entrepreneurship in the development of new ventures. The results show that the potential demand of future enterprises for big data talents can be obtained by fitting prediction sequences. Based on the Backpropagation–Autoregressive Integrated Moving Average model, the post modeling and prediction are carried out, and the coefficient 0.6235 is obtained by substituting the equation of Pearson's correlation coefficient. The analysis results suggest that the matching needs to be strengthened between the cultivation of innovative talents in universities and the demand trend of big data-related positions in enterprises. Moreover, there is a mismatch between the cultivation of innovative talents and the demand for innovative talents. Meanwhile, the mental health level of young entrepreneurs is concerned. The mental health status of young entrepreneurs is compared with the national norm data through the questionnaire survey and statistical data analysis. The results reveal that the mental health level of young entrepreneurs is significantly lower than that of the national norm, and the proportion of anxiety and depression is 29.4% and 27.5%, respectively. Considering the professional characteristics and competitive environment of young entrepreneurs, busy work and the multiple missions given by society to entrepreneurs are the major reasons for their pressure, and mental health problems are serious.

Association Between Immediacy of Citations and Altmetrics in COVID-19 Research by Artificial Neural Networks

OBJECTIVES

Both citations and Altmetrics are indexes of influence of a publication, potentially useful, but to what extent that the professional-academic citation and media-dominated Altmetrics are consistent with each other is a topic worthy of being investigated. The objective is to show their correlation.

METHODS

DOI and citation information of coronavirus disease 2019 (COVID-19) researches were obtained from the Web of Science, its Altmetric indicators were collected from the Altmetrics. Correlation between the immediacy of citation and Altmetrics of COVID-19 research was studied by artificial neural networks.

RESULTS

Pearson coefficients are 0.962, 0.254, 0.222, 0.239, 0.363, 0.218, 0.136, 0.134, and 0.505 (P < 0.01) for dimensions citation, attention score, journal impact factor, news, blogs, Twitter, Facebook, video, and Mendeley correlated with the SCI citation, respectively. The citations from the Web of Science and that from the Altmetrics have deviance large enough in the current. Altmetric score is not precise to describe the immediacy of citations of academic publication in COVID-19 research.

CONCLUSIONS

The effects of news, blogs, Twitter, Facebook, video, and Mendeley on SCI citations are similar to that of the journal impact factor. This paper performs a pioneer study for investigating the role of academic topics across Altmetric sources on the dissemination of scholarly publications.

A Clustering Algorithm for Multi-Modal Heterogeneous Big Data With Abnormal Data

The problems of data abnormalities and missing data are puzzling the traditional multi-modal heterogeneous big data clustering. In order to solve this issue, a multi-view heterogeneous big data clustering algorithm based on improved Kmeans clustering is established in this paper. At first, for the big data which involve heterogeneous data, based on multi view data analyzing, we propose an advanced Kmeans algorithm on the base of multi view heterogeneous system to determine the similarity detection metrics. Then, a BP neural network method is used to predict the missing attribute values, complete the missing data and restore the big data structure in heterogeneous state. Last, we ulteriorly propose a data denoising algorithm to denoise the abnormal data. Based on the above methods, we construct a framework namely BPK-means to resolve the problems of data abnormalities and missing data. Our solution approach is evaluated through rigorous performance evaluation study. Compared with the original algorithm, both theoretical verification and experimental results show that the accuracy of the proposed method is greatly improved.

A Tactile Method for Rice Plant Recognition Based on Machine Learning

Accurate and real-time recognition of rice plants is the premise underlying the implementation of precise weed control. However, achieving desired results in paddy fields using the traditional visual method is difficult due to the occlusion of rice leaves and the interference of weeds. The objective of this study was to develop a novel rice plant recognition sensor based on a tactile method which acquires tactile information through physical touch. The tactile sensor would be mounted on the paddy field weeder to provide identification information for the actuator. First, a flexible gasbag filled with air was developed, where vibration features produced by tactile and sliding feedback were acquired when this apparatus touched rice plants or weeds, allowing the subtle vibration data with identification features to be reflected through the voltage value of an air-pressured sensor mounted inside the gasbag. Second, voltage data were preprocessed by three algorithms to optimize recognition features, including dimensional feature, dimensionless feature, and fractal dimension. The three types of features were used to train and test a neural network classifier. To maximize classification accuracy, an optimum set of features (b (variance), f (kurtosis), h (waveform factor), l (box dimension), and m (Hurst exponent)) were selected using a genetic algorithm. Finally, the feature-optimized classifier was trained, and the actual performances of the sensor at different contact positions were tested. Experimental results showed that the recognition rates of the end, middle, and root of the sensor were 90.67%, 98%, and 96% respectively. A tactile-based method with intelligence could produce high accuracy for rice plant recognition, as demonstrated in this study.