Keyword Features Research Articles

A look at the focus shift in innovation literature due to Covid-19 pandemic

This study explores the problems related to the development of innovation research in the field of business and economics and the change in their characteristics following the coronavirus disease 2019 (COVID-19) pandemic. We compile a comprehensive bibliometric analysis of 17,277 pre-epidemic publications and 4,240 post-epidemic publications from the Web of Science. Using bibliometric methods and visualization tools, we present the changes in these publications following the COVID-19 pandemic, and identify the influential countries and regions, sources, and references, and obtain features of keywords over time. The results show that innovation research is rich in content, and involves a wide range; it has been focusing on emerging topics, such as those concerning low-carbon, innovation forms, and epidemic environments, following the COVID-19 pandemic. This study contributes to the body of knowledge on innovation, and helps to understand the features and structures of innovation research in business and economics.

A Keyword Detection and Context Filtering Method for Document Level Relation Extraction

Relation extraction (RE) is the core link of downstream tasks, such as information retrieval, question answering systems, and knowledge graphs. Most of the current mainstream RE technologies focus on the sentence-level corpus, which has great limitations in practical applications. Moreover, the previously proposed models based on graph neural networks or transformers try to obtain context features from the global text, ignoring the importance of local features. In practice, the relation between entity pairs can usually be inferred just through a few keywords. This paper proposes a keyword detection and context filtering method based on the Self-Attention mechanism for document-level RE. In addition, a Self-Attention Memory (SAM) module in ConvLSTM is introduced to process the document context and capture keyword features. By searching for word embeddings with high cross-attention of entity pairs, we update and record critical local features to enhance the performance of the final classification model. The experimental results on three benchmark datasets (DocRED, CDR, and GBA) show that our model achieves advanced performance within open and specialized domain relationship extraction tasks, with up to 0.87% F1 value improvement compared to the state-of-the-art methods. We have also designed experiments to demonstrate that our model can achieve superior results by its stronger contextual filtering capability compared to other methods.

Multitask Fine-Tuning for Passage Re-Ranking Using BM25 and Pseudo Relevance Feedback

Passage re-ranking is a machine learning task that estimates relevance scores between a given query and candidate passages. Keyword features based on the lexical similarities between queries and passages have been traditionally used for the passage re-ranking models. However, such approaches have a limitation; it is difficult to find semantic and contextual features beyond word-matching information. Recently, several studies based on neural pre-trained language models such as BERT overcome the limitations of traditional keyword-based models and they show significant performance improvements. Such ranking models have the advantage of finding the contextual features of queries and documents better than traditional keyword-based methods. However, these deep learning-based models require large amounts of data for training. Such training data is usually manually labeled with high cost, and how to utilize the data efficiently is an important issue. This paper proposes a fine-tuning method for efficient training of the neural re-ranking model. The proposed model utilizes data augmentation by simultaneously learning the ranking and MLM tasks during the fine-tuning stages. For the MLM task, different parts of a passage are masked at each training epoch. Even if only one pair of query and passage is given, the model is exposed to diverse cases with passages dynamically masked from the one. Also, the probability distribution of term importance is trained on the model.We calculate term importance weight by two novel methods using BM25 and pseudo relevance feedback. Terms are sampled and masked according to the importance weight. The ranking model learns representation based on the term weight distribution by executing the MLM task. A novel method with pseudo relevance feedback is applied for calculating term importance. It enables the neural ranking models to form representation according to feedbacks from an initial search stage. The proposed model is trained with data from the MS MARCO leaderboard for the re-ranking task. Our model achieves the state-of-the-art MRR@10 score in the leaderboard except for the ensemble-based method. In addition, our model demonstrates significant performance in three different evaluation metrics: MRR@10, Mean Rank, and Hit@(5,10,20,50).

Indonesian Online News Extraction and Clustering Using Evolving Clustering

43,000 online media outlets in Indonesia publish at least one to two stories every hour. The amount of information exceeds human processing capacity, resulting in several impacts for humans, such as confusion and psychological pressure. This study proposes the Evolving Clustering method that continually adapts existing model knowledge in the real, ever-evolving environment without re-clustering the data. This study also proposes feature extraction with vector space-based stemming features to improve Indonesian language stemming. The application of the system consists of seven stages, (1) Data Acquisition, (2) Data Pipeline, (3) Keyword Feature Extraction, (4) Data Aggregation, (5) Predefined Cluster using Automatic Clustering algorithm, (6) Evolving Clustering, and (7) News Clustering Result. The experimental results show that Automatic Clustering generated 388 clusters as predefined clusters from 3.000 news. One of them is the unknown cluster. Evolving clustering runs for two days to cluster the news by streaming, resulting in a total of 611 clusters. Evolving clustering goes well, both updating models and adding models. The performance of the Evolving Clustering algorithm is quite good, as evidenced by the cluster accuracy value of 88%. However, some clusters are not right. It should be re-evaluated in the keyword feature extraction process to extract the appropriate features for grouping. In the future, this method can be developed further by adding other functions, updating and adding to the model, and evaluating.

Tatt‐BiLSTM: Web service classification with topical attention‐based BiLSTM

AbstractWith the rapid growth of the number of Web services on the Internet, how to classify Web services correctly and efficiently become particularly important in service management tasks, such as service discovery, service selection, service ranking, and service recommendation. Existing functionality‐based service classification techniques have some drawbacks: (1) the keyword order and context information are not considered; (2) the embedding features of keywords are taken as equal importance to learn the classification model; (3) the topic number is hard to determine manually. Due to these drawbacks, the accuracy of service classification needs to be improved further. At present, deep learning techniques show the strong power in modeling complex and nonlinear function relationship. Thus, to address the problems above, this paper exploits attention mechanism to combine the local implicit state vector of Bidirectional Long Short‐Term Memory Network (BiLSTM) and the global hierarchical Dirichlet process (HDP) topic vector, and proposes a Web service classification approach with topical attention‐based BiLSTM. Specifically, BiLSTM is used to automatically learn the keyword feature representations of Web services. Then, the topic vectors of Web service documents are obtained with HDP by offline training, and topic attention mechanism is adopted to strengthen the feature representation by discriminating the importance or weight of different keywords in Web service documents. Finally, the enhanced Web service feature representation is used as the input of a softmax neural network layer to perform the classification prediction for Web services. Extensive experiments are conducted to validate the effectiveness of the proposed approach.

Sentiment analysis on newspaper article reviews: contribution towards improved rider optimization-based hybrid classifier

PurposeUp to date development in sentiment analysis has resulted in a symbolic growth in the volume of study, especially on more subjective text types, namely, product or movie reviews. The key difference between these texts with news articles is that their target is defined and unique across the text. Hence, the reviews on newspaper articles can deal with three subtasks: correctly spotting the target, splitting the good and bad content from the reviews on the concerned target and evaluating different opinions provided in a detailed manner. On defining these tasks, this paper aims to implement a new sentiment analysis model for article reviews from the newspaper.Design/methodology/approachHere, tweets from various newspaper articles are taken and the sentiment analysis process is done with pre-processing, semantic word extraction, feature extraction and classification. Initially, the pre-processing phase is performed, in which different steps such as stop word removal, stemming, blank space removal are carried out and it results in producing the keywords that speak about positive, negative or neutral. Further, semantic words (similar) are extracted from the available dictionary by matching the keywords. Next, the feature extraction is done for the extracted keywords and semantic words using holoentropy to attain information statistics, which results in the attainment of maximum related information. Here, two categories of holoentropy features are extracted: joint holoentropy and cross holoentropy. These extracted features of entire keywords are finally subjected to a hybrid classifier, which merges the beneficial concepts of neural network (NN), and deep belief network (DBN). For improving the performance of sentiment classification, modification is done by inducing the idea of a modified rider optimization algorithm (ROA), so-called new steering updated ROA (NSU-ROA) into NN and DBN for weight update. Hence, the average of both improved classifiers will provide the classified sentiment as positive, negative or neutral from the reviews of newspaper articles effectively.FindingsThree data sets were considered for experimentation. The results have shown that the developed NSU-ROA + DBN + NN attained high accuracy, which was 2.6% superior to particle swarm optimization, 3% superior to FireFly, 3.8% superior to grey wolf optimization, 5.5% superior to whale optimization algorithm and 3.2% superior to ROA-based DBN + NN from data set 1. The classification analysis has shown that the accuracy of the proposed NSU − DBN + NN was 3.4% enhanced than DBN + NN, 25% enhanced than DBN and 28.5% enhanced than NN and 32.3% enhanced than support vector machine from data set 2. Thus, the effective performance of the proposed NSU − ROA + DBN + NN on sentiment analysis of newspaper articles has been proved.Originality/valueThis paper adopts the latest optimization algorithm called the NSU-ROA to effectively recognize the sentiments of the newspapers with NN and DBN. This is the first work that uses NSU-ROA-based optimization for accurate identification of sentiments from newspaper articles.

A Corpus‐Based and Complex Computing Digital Media System for 3D Animation

In this paper, we design a corpus‐based 3D animation digital media system to improve the accuracy of 3D animation generation and realize crossplatform animation display. The corpus module extracts high‐precision data through web crawling, web cleaning, Chinese word separation, and text classification steps; the character animation generation module uses the semantic description method to expand the frame information description of the extracted data, calculates the object spatial 3D coordinates, and uses the built‐in animation execution script to generate 3D character animation; the improved digital media player module uses the improved digital media player to realize crossplatform display of 3D character animations using the improved digital media player. By constructing multidimensional character relationships and combining multiple visualization methods, the complex and multifaceted social relationship network is made available to users in an intuitive and more acceptable and understandable mode. Through a large number of user surveys, it is proved that the visual analysis method combining real social and virtual social proposed in this paper provides a more adequate and reliable basis for friend recommendation and social network analysis; the combination of multiple character relationships with geographical information and the use of visualization to describe multidimensional historical character relationships provides a new research perspective for the research and exploration of humanistic neighborhoods. The experimental results prove that the designed system can effectively read known contents and extract keywords and generate 3D animation based on keyword features, with a high accuracy rate, fast response time, small frame loss rate, and crossplatform display animation advantages.

50 Years of international journal of systems science: a review of the past and trends for the future

International Journal of Systems Science (IJSS) is an international journal in the fields of Automation & Control Systems, Computer Science, and Operations Research & Management Science. This paper is a celebration of the 50th anniversary of IJSS. A Web of Science count of 7528 documents were derived from 1970 to 2019 as data sources for the bibliometric analysis. First, the fundamental characteristics of the documents were identified from the bibliometric indicators, and features of keywords were revealed over the half century. A timeline and occurrence analyses were conducted via the software tools, through which popular keywords were visualised in different time periods. These were then grouped into four categories by a hierarchical clustering method, and emerging keywords were selected using burst detection analysis. This was followed by the Autoregressive Integrated Moving Average modelling to predict the future research trends in the four clusters as evidenced in the historical data. Documents were discussed in the recent 20 years towards guiding scholars in three aspects, i.e. problems addressed, popular methodologies, and impact in the relevant fields. Finally, a summary of the main findings was given in the conclusions.

Study on the Intelligent Selection Model of Fuzzy Semantic Optimal Solution in the Process of Translation Using English Corpus

In order to improve the accuracy and reasonableness of using English corpus for translation, a method of using English corpus to perform translation tasks based on fuzzy semantic optimal solution intelligent selection and inspired computing for wireless networks is proposed. The information extraction model using English corpus for translation is constructed, and the fuzzy semantic keyword feature directivity model of English corpus translation is established. Fuzzy semantic ontology feature registration method is used to calculate the fuzzy semantic intelligence optimal solution vector in English translation. The semantic fuzzy feature matching and adaptive subject word registration are realized in English translation. The fuzzy link relation of semantic ontology is established, and the fuzzy semantic optimal solution is obtained. The accuracy of machine translation in English corpus is improved. The experimental results show that the fuzzy semantic optimal solution has better registration performance and the feature matching degree of the subject words is higher, which improves the reasonableness and accuracy of translation in English corpus. At the same time, it provides a new idea for intelligent computation and recognition of wireless network.

Combining behavioral biometrics and session context analytics to enhance risk-based static authentication in web applications

The fragility of password-based authentication has been recognized and studied for several decades. It is an increasingly common industry practice to profile users based on their sessions context, such as IP ranges and Browser type in order to build a risk profile on an incoming authentication attempt. On the other hand, behavioral dynamics such as mouse and keyword features have been proposed in the scientific literature order to improve authentication, but have been shown most effective in continuous authentication scenarios. In this paper we propose to combine both fingerprinting and behavioral dynamics (for mouse and keyboard) in order to increase security of login mechanisms. We do this by using machine learning techniques that aim at high accuracy, and only occasionally raise alarms for manual inspection. We evaluate our approach on a dataset containing mouse, keyboard and session context information of 24 users and simulated attacks. We show that while context analysis and behavioural analysis on their own achieve around 0.7 accuracy on this dataset, a combined approach reaches up to 0.9 accuracy using a linear combination of the outcomes of the single models.

Comparative study of statistical features to detect the target event during disaster

Microblogs, such as facebook and twitter, have much attention among the users and organizations. Nowadays, twitter is more popular because of its real-time nature. People often interacted with real-time events such as earthquakes and floods through twitter. During a disaster, the number of posts or tweets is drastically increased in twitter. At the time of the disaster, detecting a target event is a challenging task. In this paper, a framework is proposed for observing the tweets and to detect the target event. For detecting the target event, a classifier is devised based on different combinations of statistical features such as the position of the keyword in a tweet, length of a tweet, the frequency of hashtag, and frequency of user mentions and the URL. From the result, it is evident that the combination of frequency of hashtag and position of keyword features provides good classification results than the other combinations of features. Hence, usage of two features, namely, frequency of hashtag and position of the earthquake keyword reduces the event's detection time. And also these two features are further helpful for detecting the sub-events which are used for filtering the tweets related to the disaster. Additionally, different classifiers such as Artificial Neural Networks (ANN), decision tree, and K-Nearest Neighbor (KNN) are compared by using these two features. However, Support Vector Machine (SVM) with linear kernel by using the combination of position of earthquake keyword and frequency of hashtag outperforms state-of-the-art methods. Therefore, SVM (linear kernel) with proposed features is applied for detecting the earthquake during disaster. The proposed algorithm is tested on Nepal earthquake and landslide datasets, 2015.

Domain-Oriented Topic Discovery Based on Features Extraction and Topic Clustering

Topic detection technology can automatically discover new topics on the Internet. This paper investigates domain-oriented feature extraction methods, and proposes a keyword feature extraction method ITFIDF-LP, a subject word feature extraction method LDA-SLP and a topic clustering model based on vector product similarity. A novel Domain-oriented Topic Discovery based on Features Extraction and Topic Clustering (DTD-FETC) model is proposed to analyze open source web of a domain and identify emerging topics in the domain in real time. This article describes a DTD-FETC system built for cyber security domain. It filters and aggregates web for specical security threat topics such as vulnerability and malware, and helps security staff respond quickly and defends against the emerging cyber threats as early as possible. The recall rate, accuracy and F1 value results of the DTD-FETC method applied to the cyber security dataset are all above 0.99.

Automated Kannada Text Summarization using Sentence Features

There is a growing requirement for the text summarization due to the difficulty of managing exponential increase of information accessible on the World Wide Web. Text summarization is a process to extract the contents in the original text to the shorter form which provides important information to the user. The summarizer presented in this paper produces the extractive summaries of Kannada text documents. The proposed summarizer system considers five features to determine the important sentences in the document. The features used are Term Frequency, Term Frequency-Inverse Sentence Frequency, Keywords feature, Sentence length and Sentence position. The value of each feature is computed and score for each sentence in the document is the average of all the feature score values. The sentences with the top scores are selected to be included in the extractive summary. The results of the proposed model are evaluated using ROUGE toolkit to measure the performance based on F-score of generated summary. Experimental studies on custom-built dataset with 50 Kannada text documents shows significantly better performance in producing extractive summaries as compared to human summaries.

“Resilience” as a policy keyword: Arts Council England and austerity

ABSTRACT This article examines how “resilience” appeared and became embedded as a keyword in Arts Council England’s (ACE) policy discourse from 2010, initially in response to the financial crisis in Britain and the government’s call for austerity. The general dynamic of what we call policy keywords here is thereby exemplified, while throwing light on Arts policy making at a specific historical juncture in Britain. Some of the features of such policy keywords are considered here: in terms of connotative ambiguities and associations, definitions, and naming or branding practices. Their distinctive purchase in ACE’s “resilience” policies is analysed in the process. The policies were designed to reduce public spending by appealing to normative agendas which, in this instance, seemed contingent on a larger and immediate impetus and were derived from the field of “ecological economics”.

Empirical comparison of text-based mobile apps similarity measurement techniques

ContextCode-free software similarity detection techniques have been used to support different software engineering tasks, including clustering mobile applications (apps). The way of measuring similarity may affect both the efficiency and quality of clustering solutions. However, there has been no previous comparative study of feature extraction methods used to guide mobile app clustering.ObjectiveIn this paper, we investigate different techniques to compute the similarity of apps based on their textual descriptions and evaluate their effectiveness using hierarchical agglomerative clustering.MethodTo this end we carry out an empirical study comparing five different techniques, based on topic modelling and keyword feature extraction, to cluster 12,664 apps randomly sampled from the Google Play App Store. The comparison is based on three main criteria: silhouette width measure, human judgement and execution time.ResultsThe results of our study show that using topic modelling, in addition to collocation-based and dependency-based feature extractors perform similarly in detecting app-feature similarity. However, dependency-based feature extraction performs better than any other in finding application domain similarity (ρ = 0.7,p − value < 0.01).ConclusionsCurrent categorisation in the app store studied does not exhibit a good classification quality in terms of the claimed feature space. However, a better quality can be achieved using a good feature extraction technique and a traditional clustering method.

Machine learning for phenotyping opioid overdose events.

To develop machine learning models for classifying the severity of opioid overdose events from clinical data. Opioid overdoses were identified by diagnoses codes from the Marshfield Clinic population and assigned a severity score via chart review to form a gold standard set of labels. Three primary feature sets were constructed from disparate data sources surrounding each event and used to train machine learning models for phenotyping. Random forest and penalized logistic regression models gave the best performance with cross-validated mean areas under the ROC curves (AUCs) for all severity classes of 0.893 and 0.882 respectively. Features derived from a common data model outperformed features collected from disparate data sources for the same cohort of patients (AUCs 0.893 versus 0.837, p value = 0.002). The addition of features extracted from free text to machine learning models also increased AUCs from 0.827 to 0.893 (p value < 0.0001). Key word features extracted using natural language processing (NLP) such as 'Narcan' and 'Endotracheal Tube' are important for classifying overdose event severity. Random forest models using features derived from a common data model and free text can be effective for classifying opioid overdose events.

Detection of Surgical Site Infection Utilizing Automated Feature Generation in Clinical Notes.

Postsurgical complications (PSCs) are known as a deviation from the normal postsurgical course and categorized by severity and treatment requirements. Surgical site infection (SSI) is one of major PSCs and the most common healthcare-associated infection, resulting in increased length of hospital stay and cost. In this work, we proposed an automated way to generate keyword features using sublanguage analysis with heuristics to detect SSI from cohort in clinical notes and evaluated these keywords with medical experts. To further valid our approach, we also applied different machine learning algorithms on cohort using automatically generated keywords. The results showed that our approach was able to identify SSI keywords from clinical narratives and can be used as a foundation to develop an information extraction system or support search-based natural language processing (NLP) approaches by augmenting search queries.

Mapping the knowledge structure of frailty in journal articles by text network analysis.

BackgroundsThis study was to understand the trends of frailty research and networking features of keywords from the academic articles focusing on frailty in the last four decades.MethodKeywords were extracted from articles (n = 6,424) retrieved from Web of Science, from 1981 to April 2016, using Bibexcel, and a social network analysis was conducted using Net Miner.ResultsThe core-keywords of research on frailty are constantly changing over the last 40 years. The keywords were tended to focus on impact in the 1980s, and moved to the determinants (i.e., malnutrition) in the 1990s and the 2000s, and in the 2010s, most of keywords were about determinants and measurement of frailty. In the early stages of frailty research, individual behaviour modifications were emphasized as intervention. Keywords with the highest degree centralities were ‘impact’ (1980s), ‘frailty’ (1990s), ‘home care’ (2000s), and ‘dementia’ (2010s). Keywords with the highest betweenness centralities were ‘model’ (1980s), ‘frailty’ (1990s), ‘chronic disease’ (2000s), and ‘malnutrition’ (2010s).DiscussionThis study provides a systematic overview of frailty knowledge development. ‘Dementia’ was found to be the keyword with the highest degree centrality, showing that studies on cognitive function are those being most actively conducted in recent decade. In the 2000s frailty research, sub-themes were sarcopenia, dementia and disability, indicating that frailty was investigated from the view of disease. In the 2010s, obesity, nutrition, prevention, evaluation, and ADL (activities of daily living) were sub-themes of the research network that focused on frailty prevention.

Question Categorization using Lexical Feature in Opini.id

This research aimed to categorize questions posted in Opini.id. N-gram and Bag of Concept (BOC) were used as the lexical features. Those were combined with Naïve Bayes, Support Vector Machine (SVM), and J48 Tree as the classification method. The experiments were done by using data from online media portal to categorize questions posted by user. Based on the experiments, the best accuracy is 96,5%. It is obtained by using the combination of Bigram Trigram Keyword (BTK) features with J48 Tree as classifier. Meanwhile, the combination of Unigram Bigram (UB) and Unigram Bigram Keyword (UBK) with attribute selection in WEKA achieves the accuracy of 95,94% by using SVM as the classifier.

KEYWORDS AS LEXICAL-SEMANTIC SUGGESTOGEN OF UKRAINIAN SERMONS

The purpose of this article is to establish the influential capacity of the key words of religious sermons. The goal requires the following tasks: to clarify the concept of "key words", to provide criteria for their selection; classify key words of religious sermons at the lexical-semantic groups; to define the phenomenon of suggestion; to highlight the suggestive potential of the key words in the genre of the sermon. The object of research are the texts of Ukrainian-language sermon as a conceptual component of religious discourse. The subject of study is lexical-semantic features of key words as a means of suggestion of religious sermons.The article used general scientific methods such as descriptive and observation of empirical material, analysis and inductive generalization of the obtained facts. The method of quantitative analysis by which set amount of keywords, the method of component analysis and contextual interpretation method allowed us to determine the specificity of the semantics of the key words in the texts of the sermons. The key words emphasize important within the Christian system of values and latent reinforce confessional model of attitude. In General, all the selected LSG of keywords can be grouped around two core the religious discourse of the concepts of divine and sinful. The semantic relationship between the key words within the limits of these poles has a wide system of semantic correlations. Verbal embodiment of such a diverse semantics gives the ability to believers that are present in Church during the sermon, it is easy to navigate categories, to determine quickly from a moral and ethical priorities and to choose behavioral tactics and strategies. That is, actually has a programming influence on the parishioners.