期刊:
Information Processing & Management,2023年60(4):103348 ISSN:0306-4573
通讯作者:
Duantengchuan Li<&wdkj&>Yan Zhang
作者机构:
[Li, Zhifei; Zhang, Yan] Hubei Univ, Sch Comp Sci & Informat Engn, Wuhan 430062, Hubei, Peoples R China.;[Zhang, Qi] Cent China Normal Univ, Sch Informat Management, Wuhan 430072, Hubei, Peoples R China.;[Zhu, Fangfang] Cent China Normal Univ, Fac Artificial Intelligence Educ, Wuhan 430079, Hubei, Peoples R China.;[Zheng, Chao; Li, Duantengchuan] Wuhan Univ, Sch Comp Sci, Wuhan 430072, Hubei, Peoples R China.
通讯机构:
[Duantengchuan Li; Yan Zhang] S;School of Computer Science, Wuhan University, Wuhan, Hubei 430072, China<&wdkj&>School of Computer Science and Information Engineering, Hubei University, Wuhan, Hubei 430062, China
作者机构:
[Cao, Shiyang] Shanxi Univ Finance & Econ, Int Exchange & Cooperat Dept, Taiyuan, Peoples R China.;[Ma, Xiao] Zhongnan Univ Econ & Law, Sch Informat & Safety Engn, Wuhan, Peoples R China.;[Yi, Ming; Zeng, Jiangfeng] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Zeng, Jiangfeng] Ctr Data Governance & Intelligent Decis Making Hub, Wuhan, Peoples R China.
通讯机构:
[Zeng, JF ] C;Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;Ctr Data Governance & Intelligent Decis Making Hub, Wuhan, Peoples R China.
关键词:
Financial sentiment analysis;Fresh and hot opinions;Temporal modeling;Fresh-hot bilinear pooling
摘要:
Financial sentiment analysis aims to extract public opinion about an institution to help financial researchers make better decisions. To predict sentiment more accurately, it is necessary for models to improve their capability to capture long-term temporal information and support multi-user interaction. However, existing methods only analyze sentiment based on one comment from a user, which fails to fully exploit the latent emotions of the public, and they lack effective temporal modeling and interaction capabilities. In this paper, we analyze a company from two perspectives to alleviate the above issues: (1) the fresh opinions can reflect timely public attitudes towards a company, while (2) the hot opinions provide the most influential views. A comprehensive exploration of fresh and hot financial sentiment can help researchers make more accurate determinations. To this end, we propose a novel financial sentiment classification framework (FSCN), that can capture temporal information and interact with the opinions of users to make a more comprehensive decision. Our approach takes into account the inherent temporal dependencies in public opinions and combines both views of information to achieve an accurate classification of financial sentiment. Specifically, the FSCN contains (1) a multi-opinion extractor to filter and extract features from massively fresh and hot opinions, respectively. (2) a fresh-hot bilinear pooling (FHBP) module to effectively fuse fresh and hot features. Additionally, to verify the effectiveness of the proposed method, we crawl data from the Internet and create a real-world public opinion dataset that consists of 79,350 comments from 837 companies. Extensive experiments demonstrate that our framework achieves state-of-the-art results on this real-world dataset and is capable of providing reliable service in the financial system. Codes will be released at https://github.com/zjfgh2015/FSCN .
作者机构:
[Liang, Han; Chen, Jincai] Huazhong Univ Sci & Technol, Wuhan Natl Lab Optoelect, Wuhan, Peoples R China.;[Chen, Jincai; Lu, Ping] Huazhong Univ Sci & Technol, Inst Nat & Math Sci, Wuhan, Peoples R China.;[Wang, Ruili; Liang, Han] Massey Univ, Inst Nat & Math Sci, Auckland, New Zealand.;[Zeng, Jiangfeng] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Zeng, Jiangfeng] Ctr Data Governance & Intelligent Decis Making Hub, Wuhan, Peoples R China.
通讯机构:
[Zeng, JF ] C;Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;Ctr Data Governance & Intelligent Decis Making Hub, Wuhan, Peoples R China.
关键词:
Audio-visual event localization;Dynamic fusion;Attention mechanism;Difference loss
摘要:
BACKGROUND: The web-based health question-and-answer (Q&A) community has become the primary and handy way for people to access health information and knowledge directly. OBJECTIVE: The objective of our study is to investigate how content-related, context-related, and user-related variables influence the answerability and popularity of health-related posts based on a user-dynamic, social network, and topic-dynamic semantic network, respectively. METHODS: Full-scale data on health consultations were acquired from the Metafilter Q&A community. These variables were designed in terms of context, content, and contributors. Negative binomial regression models were used to examine the influence of these variables on the favorite and comment counts of a health-related post. RESULTS: A total of 18,099 post records were collected from a well-known Q&A community. The findings of this study include the following. Content-related variables have a strong impact on both the answerability and popularity of posts. Notably, sentiment values were positively related to favorite counts and negatively associated with comment counts. User-related variables significantly affected the answerability and popularity of posts. Specifically, participation intensity was positively related to comment count and negatively associated with favorite count. Sociability breadth only had a significant impact on comment count. Context-related variables have a more substantial influence on the popularity of posts than on their answerability. The topic diversity variable exhibits an inverse correlation with the comment count while manifesting a positive correlation with the favorite count. Nevertheless, topic intensity has a significant effect only on favorite count. CONCLUSIONS: The research results not only reveal the factors influencing the answerability and popularity of health-related posts, which can help them obtain high-quality answers more efficiently, but also provide a theoretical basis for platform operators to enhance user engagement within health Q&A communities.
摘要:
Unsupervised sentence embedding learning is a fundamental task in natural language processing. Recently, unsupervised contrastive learning based on pre-trained language models has shown impressive performance in sentence embedding learning. This method aims to align positive sentence pairs while pushing apart negative sentence pairs to achieve semantic uniformity in the representation space. However, most previous literature leverages a random strategy to sample negative pairs, which suffers from the risk of selecting uninformative negative examples (e.g., easily distinguishable examples, anisotropic representations), thus greatly affecting the quality of learned representations. To address this issue, we propose nmCSE, a negative mining contrastive learning method for sentence embedding. Specifically, we introduce distance moderation and spatial uniformity as two properties of informative negative examples, and devise distance-based weighting and grid sampling as two strategies to preserve these properties, respectively. Our proposal outperforms the random strategy across seven semantic textual similarity datasets. Furthermore, our method can easily be adapted to other contrastive learning scenarios (e.g., vision), and does not introduce significant computational overhead.
期刊:
DIGITAL HEALTH,2023年9:20552076231208559 ISSN:2055-2076
通讯作者:
Cao, GH
作者机构:
[Gong, Hongcun; Deng, Sanhong; Wang, Hao] Nanjing Univ, Sch Informat Management, Nanjing, Peoples R China.;[Gong, Hongcun; Deng, Sanhong; Wang, Hao] Nanjing Univ, Int Joint Informat Lab, Nanjing, Peoples R China.;[Cao, Gaohui] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Cao, Gaohui; Cao, GH] Cent China Normal Univ, Sch Informat Management, 152 Luoyu Rd, Wuhan, Hubei, Peoples R China.
通讯机构:
[Cao, GH ] C;Cent China Normal Univ, Sch Informat Management, 152 Luoyu Rd, Wuhan, Hubei, Peoples R China.
关键词:
Health anxiety;health information-seeking behavior;ABC theory of emotion;rural population
摘要:
OBJECTIVE: The aim of the current study was to explore the relationship between online and offline health information-seeking behaviors, as antecedents and consequences, and health anxiety and related belief factors in rural residents. METHODS: Based on the ABC theory of emotions (ABC model), this study developed two theoretical models of the association between health anxiety and health information-seeking behavior: Placing health information-seeking behavior (both online and offline) as an outcome and antecedent, respectively, and setting three belief factors of the perceived health threat, intolerance of uncertainty, and catastrophic misinterpretations. We collected 730 self-reported data points from 20 June to 5 July 2022 for rural residents in China and empirically tested the research model and hypotheses using partial least squares-based structural equation modeling. RESULTS: The perceived health threat and intolerance of uncertainty are significant motivators of health anxiety, and health anxiety has a direct beneficial effect on both online and offline health information-seeking behaviors. Health anxiety is influenced either directly or indirectly by catastrophic misinterpretations resulting from online health information-seeking, while offline health information-seeking behavior does not contribute as strongly to health anxiety directly but mainly reinforces it through the mediating influence of catastrophic misinterpretations. CONCLUSIONS: Rural residents' health anxiety promotes their online and offline health information behaviors. And both their online and offline health information-seeking behaviors may promote health anxiety directly and through catastrophic misinterpretations. Comparing the two, online health information-seeking behaviors primarily exacerbate health anxiety through direct effects, whereas offline health information-seeking behaviors primarily affect health anxiety through catastrophic misinterpretations. We provide suggested guidelines for alleviating health anxiety and regulating health information behaviors among rural residents.
作者机构:
[Wang, Chao; Zhang, Jiaxu; Tu, Zhigang] Wuhan Univ, State Key Lab Informat Engn Surveying, Wuhan 430072, Hubei, Peoples R China.;[Xie, Wei] Cent China Normal Univ, Sch Comp, Wuhan 430079, Hubei, Peoples R China.;[Tu, Ruide] Cent China Normal Univ, Sch Informat Management, Wuhan 430079, Hubei, Peoples R China.
通讯机构:
[Chao Wang; Ruide Tu] S;State Key Laboratory of Information Engineering in Surveying, Wuhan University, Wuhan, China<&wdkj&>School Of Information Management, Central China Normal University, Wuhan, China
关键词:
Skeleton action recognition;Visual transformer;Graph-aware transformer;Velocity information of human body joints;Graph neural network
摘要:
Recently, graph convolutional networks (GCNs) play a critical role in skeleton-based human action recognition. However, most GCN-based methods still have two main limitations: (1) The semantic-level adjacency matrix of the skeleton graph is difficult to be manually defined, which restricts the perception field of GCN and limits its ability to extract the spatial–temporal features. (2) The velocity information of human body joints cannot be efficiently used and fully exploited by GCN, because GCN does not represent the correlation between the velocity vectors explicitly. To address these issues, we propose a graph-aware transformer (GAT), which can make full use of the velocity information and learn discriminative spatial–temporal motion features from the sequence of the skeleton graphs in a data-driven way. Besides, similar to the GCN-based model, our GAT also considers the prior structures of the human body including the link-aware structure and the part-aware structure. Extensive experiments on three large-scale datasets, i.e., NTU-RGB+D 60, NTU-RGB+D 120, and Kinetics-Skeleton, demonstrated that the proposed GAT obtains significant improvement compared to the GCN-based baseline for skeleton action recognition.
期刊:
Transportmetrica A: Transport Science,2023年19(2):Article: 1980131 ISSN:2324-9935
通讯作者:
Chen, Anthony
作者机构:
[Wang, Guangchao; Tong, Kebo] Cent China Normal Univ, Sch Informat Management, Wuhan 430079, Peoples R China.;[Chen, Anthony] Hong Kong Polytech Univ, Dept Civil & Environm Engn, Hong Kong, Peoples R China.;[Qi, Hang] Hubei Univ Econ, Inst Adv Studies Finance & Econ, Wuhan 430000, Peoples R China.;[Xu, Xiangdong] Tongji Univ, Key Lab Rd & Traff Engn, Shanghai 201804, Peoples R China.;[Ma, Shoufeng] Tianjin Univ, Coll Management & Econ, Tianjin 300072, Peoples R China.
通讯机构:
[Chen, Anthony] H;Hong Kong Polytech Univ, Dept Civil & Environm Engn, Hung Hom, Kowloon, Hong Kong, Peoples R China.
关键词:
Stochastic user equilibrium;least perceived travel cost;Weibit model;location parameter;relative variability
摘要:
This study investigates the impacts of the least perceived travel cost on the stochastic user equilibrium (SUE) problem. The Weibit SUE models are considered since they have a location parameter that naturally capture the least perceived travel cost. Considering a positive location parameter enhances the behavioral reality by attaching a positive lower-bound to the perceived travel cost distributions. It reduces the perception variances route-specifically and causes route-specific coefficients of variation (CVs). The CVs reduce proportionally slower for shorter routes, thus contributing to resolving the scale insensitivity issue in the Weibit SUE models. In the meantime, the route-specific CVs cause better discrimination between short and long routes in terms of relative variability; more travelers shift to the shortest route between each origin-destination pair. Numerical results confirm the analytical results regarding the effects of the least perceived travel costs and demonstrate the efficiency and robustness of the proposed solution algorithm.
摘要:
Interdisciplinary concept association discovery is a fundamental task in interdisciplinary knowledge organization. Unlike general concept association, interdisciplinary concept association mainly manifests in the correlation between fine-grained concept properties, which requires that interdisciplinary concept association discovery be explored through a fine-grained semantic association discovery tool. Existing concept association discovery methods are limited in their ability to identify interdisciplinary concept associations at fine-grained conceptual properties because they can only identify which two concepts are associated at the coarse level. To bridge this gap, we propose a method we called interdisciplinary concept association discovery based on metaphor interpretation (ICAD-MI). First, we explored the mechanism of interdisciplinary conceptual metaphor on both the cognitive and language layers, which provides a solid foundation for our method. Second, we introduced the four-step ICAD-MI method, which integrates deep learning techniques with word semantics and multidimensional contexts. We tested the ICAD-MI framework using a dataset comprising a total of 1,915 data points of interdisciplinary metaphorical expressions (IMEs) on a typical interdisciplinary conceptual metaphor Computer is a brain. Our model achieved a precision of 94.4%, a recall of 73.9%, and an F1 score of 82.9%, which outperforms the four baseline methods. Additionally, we conducted parameter analysis to further validate the effectiveness of our proposed method. The code and datasets are publicly available at: https://github.com/haihua0913/ICADMI. & COPY; 2023 Elsevier B.V. All rights reserved.
期刊:
Information Processing & Management,2023年60(4):103350 ISSN:0306-4573
通讯作者:
Li, DTC;Shi, FB
作者机构:
[Zheng, Chao; Wang, Jian; Li, Duantengchuan; Wang, Jingxiong; Li, Bing] Wuhan Univ, Sch Comp Sci, Wuhan, Peoples R China.;[Zhang, Qi] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Shi, Fobo; Shi, FB] Cent China Normal Univ, Natl Engn Res Ctr E Learning, Wuhan, Peoples R China.;[Cai, Yuefeng] ZTE Corp, Wuhan 430223, Peoples R China.;[Wang, Xiaoguang; Zhang, Zhen] Wuhan Univ, Sch Informat Management, Wuhan, Peoples R China.
通讯机构:
[Li, DTC ] W;[Shi, FB ] C;Wuhan Univ, Sch Comp Sci, Wuhan, Peoples R China.;Cent China Normal Univ, Natl Engn Res Ctr E Learning, Wuhan, Peoples R China.
摘要:
Knowledge graphs are sizeable graph-structured knowledge with both abstract and concrete concepts in the form of entities and relations. Recently, convolutional neural networks have achieved outstanding results for more expressive representations of knowledge graphs. However, existing deep learning-based models exploit semantic information from single-level feature interaction, potentially limiting expressiveness. We propose a knowledge graph embedding model with an attention-based high-low level features interaction convolutional network called ConvHLE to alleviate this issue. This model effectively harvests richer semantic information and generates more expressive representations. Concretely, the multilayer convolutional neural network is utilized to fuse high-low level features. Then, features in fused feature maps interact with other informative neighbors through the criss-cross attention mechanism, which expands the receptive fields and boosts the quality of interactions. Finally, a plausibility score function is proposed for the evaluation of our model. The performance of ConvHLE is experimentally investigated on six benchmark datasets with individual characteristics. Extensive experimental results prove that ConvHLE learns more expressive and discriminative feature representations and has outperformed other state-of-the-art baselines over most metrics when addressing link prediction tasks. Comparing MRR and Hits@1 on FB15K-237, our model outperforms the baseline ConvE by 13.5% and 16.0%, respectively.
期刊:
Transportation Research Record,2023年2677(6):530 - 547 ISSN:0361-1981
通讯作者:
Yi Xiao
作者机构:
[Yi, Ming; Xue, Xiaofei; Xiao, Yi] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Hu, Yi] Univ Chinese Acad Sci, Sch Econ & Management, Beijing, Peoples R China.
通讯机构:
[Yi Xiao] S;School of Information Management, Central China Normal University, Wuhan, People’s Republic of China
关键词:
container throughput;prediction;decomposition and ensemble;attention mechanism;artificial intelligence
摘要:
Container shipping has suffered a sharp decline since COVID-19, and risks associated with container transit will persist in the future. The decrease in container transportation has caused a ripple impact on the global supply chain. However, container throughput forecasting is both critical and complicated under the circumstances of economic uncertainty and the outbreak of the COVID-19 pandemic. A novel model propounded in this paper for container throughput forecasting to assist the port management bureau and container shipping industry integrates with the variational mode decomposition (VMD) algorithm, SARIMA technique, convolutional neural network (CNN) method, long short-term memory (LSTM) approach, and attention mechanism, among others. In this model, there are three stages: (i) data decomposition, (ii) component prediction, and (iii) ensemble output. In the first stage, the original data of the container throughput time series is decomposed into several different components using the VMD algorithm. Next, from low frequency to high frequency, each component is modeled by the corresponding prediction approach. Subsequently, the prediction results of each component generated by the previous stage are integrated into the final forecasting results by addition strategy. To enhance the prediction accuracy in the second stage, the attention mechanism is adopted in the CNN-bidirectional LSTM method. Finally, six measurement criteria, the container throughput times series at four ports, and a statistical evaluation approach are applied to comprehensively evaluate the proposed model compared with seven benchmark models. The empirical analysis demonstrates that the proposed model significantly outperforms other comparable models with regard to prediction results, level, and directional prediction accuracy.
摘要:
Cybersecurity presents non-negligible challenges for firm collaboration and supply chain viability, as information exchange among nodes introduces potential interdependent risks. How to make appropriate decisions on security investments and information exchange modes is a significant issue for supply chain members. Considering two information exchange modes: system interconnection and system independence, this study develops two game models to investigate the cybersecurity investments in a vertical supply chain composed of a retailer and n suppliers. Initial analysis shows that although firms learn investment decisions mutually in the face of a changing cybersecurity environment, suppliers always take a free ride on the efforts of retailer in two cases and the increased interdependent risks will damp nodes enthusiasm for security investments. Next, to compare the two cases, we introduce information exchange efficiency as the mediate parameter to link degree of system interconnection and proportion of information shared. We found that under the system independence mode, firms with high information-exchanging demand in the large-scale supply chain are more motivated to invest in cybersecurity. Furthermore, we extend our models to a centralised decision-making scenario. We find that security investment efficiency is greatly improved, and the free-riding behaviour of supplier is significantly reduced when systems are interconnected.
摘要:
Travel online reviews is important experience related information for understanding an inherent personality trait, novelty seeking (NS), which influences tourism motivation and the choice of tourism destinations. Manual classification of these reviews is challenging due to their high volume and unstructured nature. This paper aims to develop a classification framework and deep learning model to overcome these limitations. A multi-dimensional classification framework was created for NS personality trait that includes four dimensions synthesized from prior literature: relaxation seeking, experience seeking, arousal seeking and boredom alleviation. Based on 30 000 reviews from TripAdvisor we propose a deep learning model using Bidirectional Encoder Representations from Transformers (BERT)- Bidirectional Gated Recurrent Unit (BiGRU) to recognize NS automatically from the reviews. The classifier based on BERT- BiGRU and NS multi-dimensional scales achieved precision and F1 scores of 93.4% and 93.3% respectively, showing that NS personality trait can be relatively accurately recognized. This study also demonstrates that the classifier based on multi-dimensional NS scales can produce satisfactory results using the deep learning model. The findings also indicate that the BERT- BiGRU model achieves the best effect compared to the same kind of deep learning models. Moreover, it proves that personality traits can be automatically identified from travel reviews based on computational techniques. For practical purposes, this study provides a comprehensive classification framework for NS, which can be used in marketing and recommendation systems operating in the tourism industry.
期刊:
Journal of Information Science,2023年 ISSN:0165-5515
通讯作者:
Wu, D;Liu, XZ
作者机构:
[Dong, Jing] Cent China Normal Univ, Sch Informat Management, Wuhan, Peoples R China.;[Kang, Yangyang; Sun, Changlong] Alibaba Grp, Shanghai, Peoples R China.;[Liu, Jiawei; Wu, Dan] Wuhan Univ, Sch Informat Management, Wuhan, Peoples R China.;[Fan, Shu] Sichuan Univ, Sch Publ Adm, Sichuan, Peoples R China.;[Jin, Huchong] Indiana Univ, Luddy Sch Informat Comp & Engn, Bloomington, IN USA.
通讯机构:
[Wu, D ] W;[Liu, XZ ] I;Wuhan Univ, Sch Informat Management, 299 Bayi Rd, Wuhan 430072, Hubei, Peoples R China.;Indiana Univ Bloomington, 107 S Indiana Ave, Bloomington, IN 47405 USA.
关键词:
Active learning;annotation cost;crowdsourcing;ground truth labels;human annotations;human-centred design
摘要:
Active learning in machine learning is an effective approach to reducing the cost of human efforts for generating labels. The iterative process of active learning involves a human annotation step, during which crowdsourcing could be leveraged. It is essential for organisations adopting the active learning method to obtain a high model performance. This study aims to identify effective crowdsourcing interaction designs to promote the quality of human annotations and therefore the natural language processing (NLP)-based machine learning model performance. Specifically, the study experimented with four human-centred design techniques: highlight, guidelines, validation and text amount. Based on different combinations of the four design elements, the study developed 15 different annotation interfaces and recruited crowd workers to annotate texts with these interfaces. Annotated data under different designs were used separately to iteratively train a machine learning model. The results show that the design techniques of highlight and guideline play an essential role in improving the quality of human labels and therefore the performance of active learning models, while the impact of validation and text amount on model performance can be either positive in some cases or negative in other cases. The 'simple' designs (i.e. D1, D2, D7 and D14) with a few design techniques contribute to the top performance of models. The results provide practical implications to inspire the design of a crowdsourcing labelling system used for active learning.
摘要:
When the freshness of the firm's food is not observable, consumers may experience psychological feelings of elation or disappointment when the perceived actual freshness exceeds or falls short of their initial expectations. This paper investigates the firm's optimal selling strategies considering the consumer reference freshness effect in a fresh food supply chain. We consider two selling modes: the dual-channel retailing strategy (DCRS) mode in which the retailer sells food to consumers from store and online, and the omni-channel buy-online-and-pick-up-in-store (BOPS) mode in which the retailer additionally provides the BOPS option to consumers. Furthermore, the case of two competing suppliers is also considered. We show that on the one hand, the reference effect positively affects price strategies at a low reference freshness, and negatively affects goodwill and freshness-keeping effort in both the DCRS and BOPS modes. The BOPS mode helps to mitigate the negative effects and improve them. On the other hand, the differences found between the two modes indicate that BOPS mode is not always better than DCRS mode. Prices are higher in BOPS mode than in DCRS mode at a high reference freshness, otherwise the prices in DCRS mode are higher at a low reference freshness and with large BOPS channel size. Although BOPS mode helps to alleviate the anchoring effect, it creates higher requirements in terms of the firm's freshness-keeping efforts. If consumer perceptions of freshness as high/low are only dependent on a high/low goodwill, then the firm's profit in DCRS mode is more likely to be higher than in BOPS mode.
通讯机构:
[Ma, X ] Z;Zhongnan Univ Econ & Law, Sch Informat & Safety Engn, Wuhan, Peoples R China.
关键词:
Paper recommendation;Time-aware;Dynamic preferences;Long/short-term research interests
摘要:
With the number of scientific papers growing exponentially, recommending relevant papers for researchers has become an important and attractive research area. Existing paper recommendation methods pay more attention to the textual similarity or the citation relationships between papers. However, they generally ignore the researcher's dynamic research interests which affect the recommendation performance to a large extent. Additionally, cold start is also a serious problem in existing paper recommender systems since many researchers may have few publications, which makes the recommender systems fail to learn their preferences. In order to solve these issues, in this paper, we propose a Time-Aware Paper Recommendation (TAPRec) model, which learns researchers' dynamic preferences by encoding the long-term and short-term research interests from their historical publications. The Self-Attention method is utilized to aggregate researchers' consistent long-term research interests, while the short-term research focuses are implemented with Temporal Convolutional Networks (TCN). In addition, for researchers with few academic achievements, we combine their co-authors' dynamic preferences to solve the cold-start problem. Experiments with the DBLP dataset indicate that the proposed time-aware model performs better in the recommendation accuracy compared to the state-of-the-arts methods.
摘要:
Keeping track of translational research is essential to evaluating the performance of programs on translational medicine. Despite several indicators in previous studies, a consensus measure is still needed to represent the translational features of biomedical research at the article level. In this study, we first trained semantic representations of biomedical entities and documents (i.e., bio-entity2vec and bio-doc2vec) based on over 30 million PubMed articles. With these vectors, we then developed a new measure called Translational Progression (TP) for tracking biomedical articles along the translational continuum. We validated the effectiveness of TP from two perspectives (Clinical trial phase identification and ACH classification), which showed excellent consistency between TP and other indicators. Meanwhile, TP has several advantages. First, it can track the degree of translation of biomedical research dynamically and in real-time. Second, it is straightforward to interpret and operationalize. Third, it doesn't require labor-intensive MeSH labeling and it is suitable for big scholarly data as well as papers that are not indexed in PubMed. In addition, we examined the translational progressions of biomedical research from three dimensions (including overall distribution, time, and research topic), which revealed three significant findings. The proposed measure in this study could be used by policymakers to monitor biomedical research with high translational potential in real-time and make better decisions. It can also be adopted and improved for other domains, such as physics or computer science, to assess the application value of scientific discoveries.