Text Mining Research Articles

Star architecture in online public discourse: exploring Reddit user-generated content on the Vessel, New York, through a text analytics approach

Star architecture in online public discourse: exploring Reddit user-generated content on the Vessel, New York, through a text analytics approach

One reservoir and different views on local development: text mining analyses for the Mucharz Reservoir, Poland

Abstract In this study, we analyzed the relationship between how a community views reservoir construction and the plan for local development around it. Specifically, we focused on the social perception of the Mucharz reservoir in Poland. The research hypothesis assumed that residents’ expectations of socioeconomic development around the reservoir dam determined whether they were satisfied with the project. Fifty-nine in-depth interviews with residents of villages adjacent to the newly created dam facility were conducted. The resulting data were analyzed using text mining methods. We identified three groups of people who perceived the reservoir in different ways. Most respondents perceived the reservoir positively, mostly because of economic benefits from investments, including job creation or intensive tourism development. However, these issues do not always determine whether a local community perceives a dam project positively.

A Semantic Analysis of the User-Generated Content from Social Network Sites of Cashbuild Limited: Application of Text Mining, Machine Learning and Big Data Analytics toward the Development of a Customer-Centric Marketing Strategy

Social Network Sites (SNS) have grown rapidly over the past 15 years, which is in line with the growth of smartmobile phones. A key phenomenon of SNSs is User-Generated Content (UGC). It enables social engagements between users and their followers. UGC shares various perspectives on news, entertainment, and fun, as well as their perspectives on brands like Customer Experience (CX) and Customer Satisfaction (CS). UGC stimulates the Brand Attitudes (BA) of the readers if the user is trusted and credible. BAs influence the Purchase Intentions (PI) of the readers of UGC. Marketers like that UGC influences PI. They learn about CX and CS, which users share on the UGC. This semantic study uses Text Mining (TM) and Big Data Analysis (BDA) to extract keywords from the UGC of a building materials retail company, Cashbuild Limited (CBL) from the Republic of South Africa (RSA). The UGC was analysed using Power BI (PBI) to identify its sources and sentiments. The majority of the UGC in this study was obtained from Facebook (94.51%), followed by X.com (3.95%). The rest of the UGC was obtained from Blogs (0.79%) and Instagram (0.76%). 65.63% of the UGC was neutral, 29.74% positive, and 4.46% negative. UGC influences PI by mediating hedonic and utilitarian Brand Equity. The generation of a high-quality UGC was encouraged by CS - an outcome of CX. CX comes from the touchpoints across the CJ. Marketers want to ensure a good customer experience (CX) to achieve customer satisfaction (CS). Ultimately, CS influences future customer expectations (CE) when shared through UGC. 84 top keywords were extracted out of 16 328 hit sentences (HS) from the UGC of CBL using the MELTWATER program. The keywords were ranked to 51 frequency positions. CONCOR analysis was used to determine the connectivity and centrality of the keywords using Freeman’s coefficient and the Eigenvector value for each keyword. The results of the CONCOR analysis were used to identify 41 significant keywords based on their higher connectivity and centrality. The 41 significant keywords were matched with the 16 328 HSs. 14 keywords that matched less than 100 HSs were excluded. The remaining 27 keywords were subjected to exploratory factor analysis EFA on Statistical Program for Social Sciences (SPSS). 21 keywords were loaded on 8 factors, and the model explained 68.363% of the variance with a Kaizer Meyer Olkin (KMO) of 0.736 above 0.6. The 8 categories were compared to input from a panel of industry experts. 7 categories proposed by experts matched 7 out of 8 categories from the statistical model. 7 key themes proposed to be used in the study to develop Customer-centric Marketing Strategies (CCMS) in the building materials retail industry (BMRI) are DIY Kings, House Proud, Project Master, Renovators, Interior Image, Building Materials Inventory, and Promotions. All 5 objectives of the study were met. The study adds value to the development of CCMS learning from the UGC.

Assessment of the synergy degree of China's food safety risk governance policy tools based on text mining and machine learning methods

SummaryImplementing proactive and effective policy tools is essential for governments to manage food safety risks. Existing studies have overlooked the synergistic effects among policy tools for governing food safety risks in China. This study uses policy tool synergy to develop a quantitative model to evaluate the degree of synergy among policy tools for food safety risk governance by considering its vertical, horizontal, and temporal dimensions. Using the Top2Vec topic model and text mining techniques, we examined 558 policy documents implemented by the central and six provincial governments. The findings reveal that China's food safety risk governance system has established a collaborative mechanism, with the Administration for Market Regulation, the Department of Agriculture and Rural Affairs, and the Health Commission playing crucial roles in policy implementation. Vertical and horizontal synergies between the upper and lower levels of government and among departments at the same level are crucial for enhancing comprehensive synergy.

Application of text mining in analysing notes to financial statements: A Hungarian case

Company stakeholders must have reliable and accurate information about the companies falling into their sphere of interest. In Hungary, one of the key sources of information for company stakeholders is the financial statements and related explanations, which are included in the notes of the financial statements (notes). This study used text mining to analyse the Hungarian annual financial statements notes for 2017, 2019 and 2021. The selection of the notes was based on the proportions of each sector in the national economy. The research analysed 28,700 company notes annually, totalling 86,100 documents for the three years. The text mining and generation of the Term Frequency Matrix have performed 'quanteda' packages of the R statistical system, which incorporate the results of artificial intelligence research to enhance the efficiency of text mining. Based on the results, the contents of the notes to the financial statements appear to be a rather mixed picture in Hungary. Analysing the term frequency matrix for the 67 most common terms has revealed no significant difference between the years. However, considerable differences have been caused by size categories and sectors. The notes are statistically significant using Jaccard similarity analysis, considering the year, corporate size, and sector.

State of infection prevention knowledge among healthcare professionals in Japan: a questionnaire survey analysis using text mining

Healthcare-associated infections (HAIs) are an important issue that needs to be continuously addressed in healthcare institutions, and healthcare professionals (HCPs) are expected to strengthen educational programmes on infection prevention. Although the incidence of HAIs in Japan has been decreasing, the actual state of knowledge on infection prevention among HCPs remains unclear. To clarify the actual infection prevention knowledge of HCPs in Japan. The study participants were 1158 HCPs working in healthcare institutions with frequent contact with patients (283 doctors, 591 nurses, 115 physical therapists, 97 radiologists, and 72 medical technologists). HCPs described the infection prevention behaviours they consciously adhered to via an online self-administered questionnaire. Data were analysed by text mining. Categories were extracted from the responses to reveal HCPs' infection prevention knowledge. More than half of the participants (64.9%) were aged > 40 years, and 48.1% had over 20 years of clinical experience. The majority of the participants were nurses (51.0%), 43.9% had a bachelor's degree, and 56.6% were female. Seven categories regarding infection prevention knowledge were extracted: "performing hand hygiene and gargling," "wearing personal protective equipment," "strengthening one's immunity," "protecting oneself and patients from infection," "distinguishing clean and unclean zones," "actions to prevent transmission to others in daily life activities," and "maintaining distance from others." These results suggest that most HCPs working in healthcare settings in Japan prioritize and adhere to standard precautionary measures, and that the low incidence of HAIs may be influenced by perceptions of the knowledge of "handwashing and gargling" among HCPs.

Combating corruption on the frontlines: analyzing penalties for street-level bureaucrats in China

Penalties for corrupt acts are a vital component of integrity management systems within bureaucratic organizations, yet systematic inquiry into specific interventions is lacking. Drawing from the principal–supervisor–agent model, this article explores the criteria for imposing penalties on corrupt bureaucrats at the street level. We examine the impact of various factors on the severity of penalties for corruption, highlighting the influence of top-down anticorruption reform and local enforcement capacity. Utilizing text mining techniques, we build a comprehensive dataset containing 4025 cases of punishing corruption among grassroots officials in China from 2015 to 2023 and use multilevel analysis to explore these dynamics. Findings suggest that anticorruption reform directly led to an increase in the severity of penalties, but with substantial variation across practices. In particular, the analysis reveals a positive correlation between enforcement capacity and the severity of penalties. Moreover, while the relationship between harm and penalty severity represents a generalizable observation, significant disparities exist in the penalties imposed across various types of corruption. This study contributes to understanding the dynamics of multiple punitive measures and the factors leading to penalties imposed on corrupt street-level bureaucrats.

Improving dictionary-based named entity recognition with deep learning.

Dictionary-based named entity recognition (NER) allows terms to be detected in a corpus and normalized to biomedical databases and ontologies. However, adaptation to different entity types requires new high-quality dictionaries and associated lists of blocked names for each type. The latter are so far created by identifying cases that cause many false positives through manual inspection of individual names, a process that scales poorly. In this work, we aim to improve block list s by automatically identifying names to block, based on the context in which they appear. By comparing results of three well-established biomedical NER methods, we generated a dataset of over 12.5 million text spans where the methods agree on the boundaries and type of entity tagged. These were used to generate positive and negative examples of contexts for four entity types (genes, diseases, species, and chemicals), which were used to train a Transformer-based model (BioBERT) to perform entity type classification. Application of the best model (F1-score = 96.7%) allowed us to generate a list of problematic names that should be blocked. Introducing this into our system doubled the size of the previous list of corpus-wide blocked names. In addition, we generated a document-specific list that allows ambiguous names to be blocked in specific documents. These changes boosted text mining precision by ∼5.5% on average, and over 8.5% for chemical and 7.5% for gene names, positively affecting several biological databases utilizing this NER system, like the STRING database, with only a minor drop in recall (0.6%). All resources are available through Zenodo https://doi.org/10.5281/zenodo.11243139 and GitHub https://doi.org/10.5281/zenodo.10289360.

A study on the impact of investor sentiment on individual stocks based on text mining

This paper investigates the influence of investor sentiment on individual stock prices through text mining. By reviewing relevant theories and literature, we utilized web crawler technology to collect comment data from the Eastmoney Stock Bar, constructing a comprehensive index of investor sentiment and an index of investor sentiment for individual stocks. Employing the VAR model and Granger causality test, we analyzed the correlation and dynamic relationship between investor sentiment and stock price fluctuations. The study found that stocks of different types of financial institutions react differently to investor sentiment, and that investor sentiment significantly affects stock prices. This paper puts forward corresponding policy recommendations, which can help market investors make scientific decisions and promote the healthy development of the market.

The Analysis of Impact of Metaverse on the FinTech Industry Based on Text Mining and Association Rules Mining

The Analysis of Impact of Metaverse on the FinTech Industry Based on Text Mining and Association Rules Mining

An Analysis of trends in Japanese-Korean Contrastive Research Using Text Mining

An Analysis of trends in Japanese-Korean Contrastive Research Using Text Mining

Analysis of Media Coverage on Teen Parents Using Text Mining

Analysis of Media Coverage on Teen Parents Using Text Mining

Analysis of Changes in Learning Objectives for Educational Evaluation Courses Using Text Mining

Analysis of Changes in Learning Objectives for Educational Evaluation Courses Using Text Mining

텍스트마이닝을 활용한 도심 드론 물류 서비스 수용성에 대한 연구: 유튜브 댓글을 활용하여

In modern society, logistics systems need new innovations due to urban traffic congestion and environmental issues. Drones, or unmanned aerial vehicles, are gaining attention for improving urban logistics efficiency. However, drone delivery faces issues such as costs, regulations, and noise. This study analyzes consumer perceptions of drone delivery services using text mining technology. A total of 16,695 comments from 587 YouTube videos containing the keyword ‘drone delivery’ were analyzed through frequency analysis, topic modeling, and sentiment analysis. The results provided insights into consumer opinions, attitudes, concerns, and preferences regarding drone logistics. Major issues identified include potential accidents, noise, and high costs. Proposed solutions include enhancing safety, developing noise reduction technologies, reducing costs, and achieving social consensus. This study offers insights for improving drone delivery services and developing strategies that reflect consumer demands for future implementations.

Analyzing Online Consumer Reviews on Seafood Products Using Multifaceted Text Mining Techniques

Analyzing Online Consumer Reviews on Seafood Products Using Multifaceted Text Mining Techniques

A Review of Sentiment Analysis in Social Media Perspectives

The widespread use of the Internet and social media platforms has led to an increase in the number of individuals who declare their feelings publicly. Therefore, sentiment analysis systems have proceeded because of their crucial role in determining the personal opinions of users. This is can greatly influence the decision-making process in various fields. To create a robust and reliable sentiment analysis system, it was necessary to apply techniques capable of dealing with these scattered opinions. Natural language processing techniques are commonly used to extract information from unstructured text data published by humans. The comments and posts in social media platforms are often ignore the grammar rules and sentence structure. This is resulting in many ambiguities in lexical, syntactic, and semantic aspects. As a result, researchers have developed different methods for text mining and defining real information. This survey aims to study the different methods used in sentiment analysis filed. We discussed two common models of classification, including the vocabulary-based model and the supervision-based approach.

User Perception and Evaluation of Public Tennis Courts through Text Mining Analysis of Google Maps Reviews

User Perception and Evaluation of Public Tennis Courts through Text Mining Analysis of Google Maps Reviews

지속가능한 도시로서의 스마트 도시재생을 위한 리질리언스 정책 방향에 관한 연구

The purpose of this study is to find Policy measures that can help smart cities contribute to urban regeneration as sustainable safe cities. This is a policy direction to move toward a sustainable city by combining smart cities and urban regeneration, which are spreading as models of future safe cities. After analyzing the trends of existing papers using text mining techniques, the concept of disaster resilience was grafted to smart city regeneration. The basic elements that constitute the qualitative criteria of resilience were used as the study method for building a sustainable city. Results, the policy direction of resilience construction for a ‘sustainable urban regeneration-type smart city’ was presented in three aspects. First, it’s constructing resilience capacity in smart urban regeneration plans. The qualitative criteria for resilience capacity were presented as four: Resourcefulness Robustness, Rapidity, and Redundancy. The four elements of disaster preparedness resilience capacity can be used as qualitative criteria for resilience required for smart urban regeneration as a safe city. The Second is the construction of the ‘decision support system for resilience’ that is needed for smart urban regeneration plans. For this, seven types of infrastructure were presented. These infrastructural interrelationships form the basis for the operation of the ‘decision support system for resilience’. Third, the establishment of ‘Privacy by Design as a smart resilience’ is necessary for the preemptive protection of personal information in smart urban regeneration. Privacy protection is essential for a modern safe city. In conclusion, we hope that the policy directions of the above three aspects will be linked to legislation and contribute to the activation of urban regeneration-type smart city as a sustainable and safe city in the future.

CarD-T: Interpreting Carcinomic Lexicon via Transformers.

The identification and classification of carcinogens is critical in cancer epidemiology, necessitating updated methodologies to manage the burgeoning biomedical literature. Current systems, like those run by the International Agency for Research on Cancer (IARC) and the National Toxicology Program (NTP), face challenges due to manual vetting and disparities in carcinogen classification spurred by the volume of emerging data. To address these issues, we introduced the Carcinogen Detection via Transformers (CarD-T) framework, a text analytics approach that combines transformer-based machine learning with probabilistic statistical analysis to efficiently nominate carcinogens from scientific texts. CarD-T uses Named Entity Recognition (NER) trained on PubMed abstracts featuring known carcinogens from IARC groups and includes a context classifier to enhance accuracy and manage computational demands. Using this method, journal publication data indexed with carcinogenicity & carcinogenesis Medical Subject Headings (MeSH) terms from the last 25 years was analyzed, identifying potential carcinogens. Training CarD-T on 60% of established carcinogens (Group 1 and 2A carcinogens, IARC designation), CarD-T correctly to identifies all of the remaining Group 1 and 2A designated carcinogens from the analyzed text. In addition, CarD-T nominates roughly 1500 more entities as potential carcinogens that have at least two publications citing evidence of carcinogenicity. Comparative assessment of CarD-T against GPT-4 model reveals a high recall (0.857 vs 0.705) and F1 score (0.875 vs 0.792), and comparable precision (0.894 vs 0.903). Additionally, CarD-T highlights 554 entities that show disputing evidence for carcinogenicity. These are further analyzed using Bayesian temporal Probabilistic Carcinogenic Denomination (PCarD) to provide probabilistic evaluations of their carcinogenic status based on evolving evidence. Our findings underscore that the CarD-T framework is not only robust and effective in identifying and nominating potential carcinogens within vast biomedical literature but also efficient on consumer GPUs. This integration of advanced NLP capabilities with vital epidemiological analysis significantly enhances the agility of public health responses to carcinogen identification, thereby setting a new benchmark for automated, scalable toxicological investigations.

텍스트 마이닝 기반 고등학교 일본어 교육에 적합한 일본 대중음악 선별 연구

Language training with songs is not only effective in engaging learners, but also in improving their language skills. While a lot of research on the topic of English education using pop songs have been conducted, few have tried and tested the effectiveness of teaching Japanese using Japanese pop songs. This study aims to propose criteria for selecting songs suitable for language teaching, considering the educational goals set forth in the Japanese language curriculum. To this end, this study collected Japanese songs from domestic music sites and analyzed 634 songs. Next, this study set the criteria for selecting appropriate songs from a grammar and vocabulary perspective. Specifically, 1) the proportion of words included in the high school Japanese curriculum, 2) the number of verbs, and 3) the proportion and the variety of verb conjugations were introduced as selection criteria. In the end, three songs were selected that met a certain level of criteria for each of these criteria. This study is significant in that it proposes a method for selecting songs suitable for Japanese language education by utilizing text mining techniques.