Text Mining Model Research Articles

A text mining and machine learning study on the trends of and dynamics between collective action and mental health in politically polarized online environments

Social media and online forums play an increasingly important role in the mobilization of collective action. This study examined how the discussion of collective actions impacts the expression of psychological distress in politically polarized online environments. We used text mining and machine learning models to analyze 39,487,911 user-generated comments during the 2019 social unrest in Hong Kong on two online forums frequented by anti-government (Lihkg.com) and pro-government (Discuss.com.hk) netizens. Results from time-series models yielded two main findings. First, there was a time-lagged association between the discussion of protest and the mention of psychological distress on both forums. Second, on Discuss.com.hk but not Lihkg.com, fewer comments containing psychological distress were created on days with offline protests (especially on days with violent conflicts) than days without. Together, these findings suggest that politically polarizing environments contribute to psychological distress.

Impact of Green Features on Rental Value of Residential Properties: Evidence from South Africa

In recent years, scholars have called for an increase in the usage of green features in the built environment to address climate change issues. Governments across the developed world are implementing legislation to support this increased uptake. However, little is known about how the inclusion of green features influences the rental value of residential properties located in developing countries. Data on 389 residential properties were extracted and collected from a webpage. Text mining and machine learning models were used to evaluate the impact of green features on the rental value of residential properties. The results indicated that floor area, number of bathrooms, and availability of furniture are the top three attributes affecting the rental value of residential properties. The random forest model generated better predictions when compared with other modelling techniques. It was also observed that green features are not the most common words mentioned in rental adverts for residential properties. The results suggest that green features add limited value to residential properties in South Africa. This finding suggests that there is a need for stakeholders to create and implement policies targeted at incentivising the inclusion of green features in existing and new residential properties in South Africa.

Evaluating public opinions: informing public health policy adaptations in China amid the COVID-19 pandemic

Public concern regarding safety policies serious consequences is anticipated to persist over an extended duration. A study examining a case of rapid public health policy adaptation in China during the COVID-19 epidemic was conducted by gathering public opinion data from major social media platforms. A systematic approach to comprehend public opinion was developed. Five fundamental elements and four dimensions were delineated. An indicator system was established utilizing the K-means text clustering model. Public prediction, expectation, and their evolution underlying public concern were elucidated employing TF–IDF text mining models. The HMM elucidated the way public opinion influences policy adjustments. The findings underscore that public concern regarding enduring events undergoes temporal shifts, mirroring the evolution of public opinion towards policy. Public opinion aroused by both the original event and derived events collaboratively influence policy adjustments. In China, public opinion serves as a mechanism for policy feedback and oversight; notably, negative public sentiment plays a pivotal role in expediting policy transitions. These findings aid in refining policies to mitigate emergencies through a feedback loop, thereby averting the emergence of safety risks such as social unrest prompted by public opinion.

Understanding Food Waste in Bulk in Incheon: Based on Naver Blog Big Data

Purpose: This study aims to provide insight into the contemporary food waste issue for the better management of food waste in large quantities in Incheon, one specific region in Korea, based on the analysis of the Naver Blog corpus, employing analysis techniques including text mining and LDA topic modeling.
 Design/methodology/approach: In order to achieve the aforementioned objectives, the current study has employed the R program for the analysis (e.g., TF, tf-idf, correlation, and LDA) of 868 Naver Blog posts which included information about food waste, and/or garbage produced in bulk by foodservice operators in the context of Incheon.
 Findings: The frequently addressed keywords in the dataset include food, waste, garbage, processor, workplace, food waste dewaterer, microorganism, pulverizer, plastic, compressor, restaurant, sink, cafeteria, and bean sprouts. The correlation analysis demonstrated that waste is largely generated by food material, and food material is closely related to specific types of private companies (e.g., cafeterias) and public places (e.g., military bases, prisons, hospi-tals). The LDA identified three topics: the implications for food waste produced by the workplace, recent equipment and technologies used for food waste processing, and effect of waste on the environment and call for remedies.
 Research limitations/implications: While this study has shed light on contemporary issues in relation to food waste in Incheon, it is suggested that the involved parties in the waste management industry pay more attention to the development of effective waste management strategies by hospitality operators in a specific region.
 Originality/value: This study responds to a lack of understanding underpinning foodservice operators who produce a large quantity of food waste in Incheon, albeit much attention has been paid to some recent research on food waste.

Understanding global research trends in the control and prevention of infectious diseases for children: Insights from text mining and topic modeling.

The emergence of novel infectious diseases has amplified the urgent need for effective prevention strategies, especially ones targeting vulnerable populations such as children. Factors such as the high incidence of both emerging and existing infectious diseases, delays in vaccinations, and routine exposure in communal settings heighten children's susceptibility to infections. Despite this pressing need, a comprehensive exploration of research trends in this domain remains lacking. This study aims to address this gap by employing text mining and modeling techniques to conduct a comprehensive analysis of the existing literature, thereby identifying emerging research trends in infectious disease prevention among children. A cross-sectional text mining approach was adopted, focusing on journal articles published between January 1, 2003, and August 31, 2022. These articles, related to infectious disease prevention in children, were sourced from databases such as PubMed, CINAHL, MEDLINE (Ovid), Scopus, and Korean RISS. The data underwent preprocessing using the Natural Language Toolkit (NLTK) in Python, with a semantic network analysis and topic modeling conducted using R software. The final dataset comprised 509 journal articles extracted from multiple databases. The study began with a word frequency analysis to pinpoint relevant themes, subsequently visualized through a word cloud. Dominant terms encompassed "vaccination," "adolescent," "infant," "parent," "family," "school," "country," "household," "community," "HIV," "HPV," "COVID-19," "influenza," and "diarrhea." The semantic analysis identified "age" as a key term across infection, control, and intervention discussions. Notably, the relationship between "hand" and "handwashing" was prominent, especially in educational contexts linked with "school" and "absence." Latent Dirichlet Allocation (LDA) topic modeling further delineated seven topics related to infectious disease prevention for children, encompassing (1) educational programs, (2) vaccination efforts, (3) family-level responses, (4) care for immunocompromised individuals, (5) country-specific responses, (6) school-based strategies, and (7) persistent threats from established infectious diseases. The study emphasizes the indispensable role of personalized interventions tailored for various child demographics, highlighting the pivotal contributions of both parental guidance and school participation. The study provides insights into the complex public health challenges associated with preventing and managing infectious diseases in children. The insights derived could inform the formulation of evidence-based public health policies, steering practical interventions and fostering interdisciplinary synergy for holistic prevention strategies.

Characteristics of students’ learning behavior preferences — an analysis of self-commentary data based on the LDA model

How to better grasp students’ learning preferences in the environment of rapid development of engineering and science and technology so as to guide them to high-quality learning is one of the important research topics in the field of educational technology research today. In order to achieve this goal, this paper utilizes the LDA (Latent Dirichlet Allocation) model for text mining of the survey results on the basis of a survey on students’ self-perception evaluation. The results show that the LDA model is capable of extracting terms from text, fuzzy identifying groups of students at different levels and presenting potential logical relationships between the groups, and further analyzing the learning preferences of students at different levels for IT courses. Based on the student’s learning needs, this paper proposes recommendations for developing students’ learning effectiveness. The LDA method proposed in this paper is a feasible and effective method for assessing students’ learning dynamics as it generates cognitive content about students’ learning and allows for the timely discovery of students’ learning expectations and cutting-edge dynamics.

Navigating the new normal: Redefining N95 respirator design with an integrated text mining and quality function deployment-based optimization model

The global surge in demand during the pandemic led to a proliferation of substandard surgical N95 respirators, particularly in the e-commerce market, due to a lack of regulatory standards. In the post-pandemic era, the essential role of these respirators in protecting against airborne hazards and infectious agents persists. Addressing the need for standardized design and manufacturing processes, this research proposes a three-phased integrated decision-making framework. Firstly, text-mining algorithms analyze online consumer reviews to identify critical characteristics of surgical N95 respirators. Subsequently, a spherical-fuzzy-based quality function deployment assesses manufacturing capabilities in alignment with consumer requirements. The final phase utilizes Mixed Integer Non-linear Programming to optimize utility and select the most effective design, considering both technical specifications and consumer needs. The proposed framework is validated through a case study, revealing technical characteristics crucial for organizations in product development. The findings not only offer insights into the technical aspects of N95 respirators but also recommend essential product characteristic tests during manufacturing and certification. This information is vital for educating consumers and guiding companies in the effective use and development of Surgical N95 respirators in the post-pandemic landscape, addressing the challenges posed by inferior and counterfeit products.

텍스트 마이닝을 활용한 국내 삶의 의미 연구 동향 분석

This study aims to examine the trends of research on meaning in life and provide basic data through text mining research methods. 368 academic articles related to the meaning of life were analyzed. In addition to text mining, topic modeling was used. The results of the analysis showed that, first, studies related to the meaning of life were published in 142 academic journals such as the Korean Youth Association and the Korean Counseling Association, and were studied in various fields. Second, keyword frequency analysis showed that meaning, research, result, and relationship were keywords. Third, the topic modeling analysis showed that the topics were: enhancing life satisfaction, resolving negative emotions and experiencing positive emotions, improving emotional efficacy and career adaptability, enhancing social skills, therapeutic characteristics and effects, and program development. This study is significant in that it is an attempt to present research trends related to life meaning and analyze them through text mining and topic modeling techniques, and provides limitations and suggestions along with discussion.

Modest performance of text mining to extract health outcomes may be almost sufficient for high-quality prognostic model development

Background:Across medicine, prognostic models are used to estimate patient risk of certain future health outcomes (e.g., cardiovascular or mortality risk). To develop (or train) prognostic models, historic patient-level training data is needed containing both the predictive factors (i.e., features) and the relevant health outcomes (i.e., labels). Sometimes, when the health outcomes are not recorded in structured data, these are first extracted from textual notes using text mining techniques. Because there exist many studies utilizing text mining to obtain outcome data for prognostic model development, our aim is to study the impact of the text mining quality on downstream prognostic model performance. Methods:We conducted a simulation study charting the relationship between text mining quality and prognostic model performance using an illustrative case study about in-hospital mortality prediction in intensive care unit patients. We repeatedly developed and evaluated a prognostic model for in-hospital mortality, using outcome data extracted by multiple text mining models of varying quality. Results:Interestingly, we found in our case study that a relatively low-quality text mining model (F1 score ≈ 0.50) could already be used to train a prognostic model with quite good discrimination (area under the receiver operating characteristic curve of around 0.80). The calibration of the risks estimated by the prognostic model seemed unreliable across the majority of settings, even when text mining models were of relatively high quality (F1 ≈ 0.80). Discussion:Developing prognostic models on text-extracted outcomes using imperfect text mining models seems promising. However, it is likely that prognostic models developed using this approach may not produce well-calibrated risk estimates, and require recalibration in (possibly a smaller amount of) manually extracted outcome data.

Advanced Sampling Technique in Radiology Free-Text Data for Efficiently Building Text Mining Models by Deep Learning in Vertebral Fracture.

This study aims to establish advanced sampling methods in free-text data for efficiently building semantic text mining models using deep learning, such as identifying vertebral compression fracture (VCF) in radiology reports. We enrolled a total of 27,401 radiology free-text reports of X-ray examinations of the spine. The predictive effects were compared between text mining models built using supervised long short-term memory networks, independently derived by four sampling methods: vector sum minimization, vector sum maximization, stratified, and simple random sampling, using four fixed percentages. The drawn samples were applied to the training set, and the remaining samples were used to validate each group using different sampling methods and ratios. The predictive accuracy was measured using the area under the receiver operating characteristics (AUROC) to identify VCF. At the sampling ratios of 1/10, 1/20, 1/30, and 1/40, the highest AUROC was revealed in the sampling methods of vector sum minimization as confidence intervals of 0.981 (95%CIs: 0.980-0.983)/0.963 (95%CIs: 0.961-0.965)/0.907 (95%CIs: 0.904-0.911)/0.895 (95%CIs: 0.891-0.899), respectively. The lowest AUROC was demonstrated in the vector sum maximization. This study proposes an advanced sampling method, vector sum minimization, in free-text data that can be efficiently applied to build the text mining models by smartly drawing a small amount of critical representative samples.

Cereal-legume intercropping: a smart review using topic modelling.

Over the last decade, there has been a growing interest in cereal-legume intercropping for sustainable agriculture. As a result numerous papers, including reviews, focus on this topic. Screening this large amount of papers, to identify knowledge gaps and future research opportunities, manually, would be a complex and time consuming task. Bibliometric analysis combined with text mining and topic modelling, to automatically find topics and to derive a representation of intercropping papers as a potential solution to reduce the workload was tested. Both common (e.g. wheat and soybean) as well as underutilized crops (e.g. buckwheat, lupin, triticale) were the focus of this study. The corpus used for the analysis was retrieved from Web of Science and Scopus on 5th September 2022 and consisted of 4,732 papers. The number of papers on cereal-legume intercropping increased in recent years, with most studies being located in China. Literature mainly dealt with the cereals maize and wheat and the legume soybean whereas buckwheat and lupin received little attention from academic researchers. These underutilized crops are certainly interesting to be used as intercropping partners, however, additional research on optimization of management and cultivar's choice is important. Yield and nitrogen fixation are the most commonly studied traits in cereal-legume intercropping. Last decade, there is an increasing interest in climate resilience, sustainability and biodiversity. Also the term "ecosystem services" came into play, but still with a low frequency. The regulating services and provisioning services seem to be the most studied, in contrast terms related to potential cultural services were not encountered. In conclusion, based on this review several research opportunities were identified. Minor crops like lupin and buckwheat need to be evaluated for their role as intercropping partners. The interaction between species based on e.g. root exudates needs to be further unraveled. Also diseases, pests and weeds in relation to intercropping deserve more attention and finally more in-depth research on the additional benefits/ecosystem services associated with intercropping systems is necessary.

A study on kansei attraction of products’ online reviews by using text mining and kano model

A study on kansei attraction of products’ online reviews by using text mining and kano model

Quantitative Evaluation of Domestic Cultural and Tourism Integration Policies Based on PMC Index Modeling

Abstract This paper constructs a text mining and quantitative evaluation model for cultural and tourism integration policies based on LDA and PMC index models. We find research samples of cultural and tourism integration policies through relevant websites and databases and use the LDA model to mine high-frequency words in the policies and make a high-frequency glossary. The quantitative evaluation index system of cultural and tourism policies is constructed according to the glossary and the definitions of policy evaluation by authoritative scholars, and then the scores of each policy are calculated based on the PMC formula, and the PCM surface diagram is drawn for quantitative evaluation. The results show that 57.14% of the 63 policies belong to the excellent level, and most of the policies have high scores in terms of policy timeliness, subject, tool, field, evaluation, and function, ranging from 0.5 to 0.8. Thirteen policies were poorly rated in terms of policy nature, with only 0.333-05, and there is much room for improvement. Based on the results of the study, an optimization concept focusing on optimizing the nature of policies is proposed to improve the problems existing in the current policy formulation of cultural tourism integration.

Bibliometric analysis of literature on social media trends during the COVID-19 pandemic

PurposeThis study employs bibliometric analysis to map the research landscape of social media trending topics during the COVID-19 pandemic. The authors aim to offer a comprehensive review of the predominant research organisations and countries, key themes and favoured research methodologies pertinent to this subject.Design/methodology/approachThe authors extracted data on social media trending topics from the Web of Science Core Collection database, spanning from 2009 to 2022. A total of 1,504 publications were subjected to bibliometric analysis, utilising the VOSviewer tool. The study analytical process encompassed co-occurrence, co-authorship, citation analysis, field mapping, bibliographic coupling and co-citation analysis.FindingsInterest in social media research, particularly on trending topics during the COVID-19 pandemic, remains high despite signs of the pandemic stabilising globally. The study predominantly addresses misinformation and public health communication, with notable focus on interactions between governments and the public. Recent studies have concentrated on analysing Twitter user data through text mining, sentiment analysis and topic modelling. The authors also identify key leading organisations, countries and journals that are central to this research area.Originality/valueDiverging from the narrow focus of previous literature reviews on social media, which are often confined to particular fields or sectors, this study offers a broad view of social media's role, emphasising trending topics. The authors demonstrate a significant link between social media trends and public events, such as the COVID-19 pandemic. The paper discusses research priorities that emerged during the pandemic and outlines potential methodologies for future studies, advocating for a greater emphasis on qualitative approaches.Peer reviewThe peer-review history for this article is available at: https://publons.com/publon/10.1108/OIR-05-2023-0194.

A GOLD MINE OF TEXT TEXT MINING VERSUS MANUAL CODING METHODS REGARDING QUALITY OF CARE IN LONG-TERM CARE FOR OLDER ADULTS

Abstract In long-term care for older adults, large amounts of text are collected that cover the quality of care, such as transcribed interviews, medical records, and text in electronic health records. Researchers currently analyze textual data manually to gain insights, which is a time-consuming process. Text mining (TM) could provide a solution, as this methodology can analyze large amounts of text automatically. Therefore, this study aims to compare TM to the gold standard of manual coding regarding sentiment analysis and thematic content analysis. Data was collected from interviews with residents (n=21), family (n=20) and care professionals (n=20). TM models were developed and compared to the manual approach. The results of the manual and TM approach were evaluated in three ways: accuracy, consistency and expert feedback. Accuracy showed how similar the approaches are. Consistency showed if an individual approach finds the same themes in similar text segments. Expert feedback was used to show the perceived correctness of the TM approach. The accuracy analysis showed that 81.8% of text segments receive the same sentiment code in both approaches and 83.7% of text segments receive the same thematic codes. Interviews coded by TM had a higher consistency compared to those coded manually. Expert feedback showed that the TM had a limited understanding of the context. However, it also showed certain inconsistencies of manual coding. The current study has shown that TM can be an effective tool for quickly and accurately identifying sentiment and thematic content in large amounts of textual data.

Hotel Reviews Analysis Using Machine Learning Algorithms and Text Mining Model

Hotel Reviews Analysis Using Machine Learning Algorithms and Text Mining Model

“The Times They Are A-Changin' “ – research trends and perspectives of reindeer pastoralism – A review using text mining and topic modelling

A literature review on reindeer pastoralism was performed to assess research trends within this multifaceted field. A machine learning topic modelling approach was applied on abstracts retrieved online. The methodology offers a computational tool to find relevant themes by capturing meaningful structure among large collections of documents. Ten topics were identified. The three fastest expanding topics were related to “Space use”, “Climate change” and “Socio-cultural” aspects. They are interrelated and reflect the main challenges herders and reindeer are facing as their land use competes with an expanding resource extraction industry, arguably for facilitating the Green shift. Climate change thus indirectly influence the herders’ rights to land and resources and form of life. The “Socio-cultural” topic outnumbers the other topics during the last decade. The dominating natural science argument encompasses four topics: “Space use”, “Climate change”, “Herd productivity” and “Grazing and Vegetation”. Research across disciplines is still in its infancy, and an integration between social and ecological issues through a multidisciplinary approach demands for a deeper understanding of the future challenges faced by reindeer pastoralism.

Comparing text mining and manual coding methods: Analysing interview data on quality of care in long-term care for older adults.

In long-term care for older adults, large amounts of text are collected relating to the quality of care, such as transcribed interviews. Researchers currently analyze textual data manually to gain insights, which is a time-consuming process. Text mining could provide a solution, as this methodology can be used to analyze large amounts of text automatically. This study aims to compare text mining to manual coding with regard to sentiment analysis and thematic content analysis. Data were collected from interviews with residents (n = 21), family members (n = 20), and care professionals (n = 20). Text mining models were developed and compared to the manual approach. The results of the manual and text mining approaches were evaluated based on three criteria: accuracy, consistency, and expert feedback. Accuracy assessed the similarity between the two approaches, while consistency determined whether each individual approach found the same themes in similar text segments. Expert feedback served as a representation of the perceived correctness of the text mining approach. An accuracy analysis revealed that more than 80% of the text segments were assigned the same themes and sentiment using both text mining and manual approaches. Interviews coded with text mining demonstrated higher consistency compared to those coded manually. Expert feedback identified certain limitations in both the text mining and manual approaches. While these analyses highlighted the current limitations of text mining, they also exposed certain inconsistencies in manual analysis. This information suggests that text mining has the potential to be an effective and efficient tool for analysing large volumes of textual data in the context of long-term care for older adults.

A new multifractal-based deep learning model for text mining

Text mining is indispensable and heavily relied upon in artificial intelligence systems. Nevertheless, a disregard for the complexity of language expression and the challenges like gradient vanishing within prevailing activation functions, shackle the potential of text mining. In this paper, we introduce a new text mining model that deftly marries the concept of fractals to delve into the language's complexity. It employs our proposed multifractal approach to alleviate potential noise, simultaneously harnessing the potency of our proposed activation function to amplify the neural network's performance. The success on experiments anchored in real-world technical reports covering the recognition of technical term and classification of hazard events, stands as a testament to our endeavors. This research venture not only expands our understanding of text mining but also opens new horizons for information processing across various domains.

Cats' and dogs' welfare: text mining and topics modeling analysis of the scientific literature.

Animal welfare is a field with increasing significance and has been raising huge concerns of the public and the political stage. Cats and dogs possess an important role in human life, but their welfare is not always secured from a legal aspect. This review aimed to describe the evolution and geographical distribution of "cats and dogs" and "puppies and kittens" welfare literature over the last 40 years, distinguish the main research topics studied and highlight gaps in knowledge. A search using Scopus® was performed with different search strings and predetermined filters as time range, language, and subject area. A total of 2,725 scientific literature records were retrieved but only the ones that referred to cats and dogs' welfare aspects were retained. The final 1,775 records were processed through descriptive statistics, and text mining and topic analysis procedures were performed on their titles and abstracts. The results showed that the number of studies has been increasing, especially in Europe and North America. "Shelter" was the most frequent word, followed by "behavior," "owner" and "adopt." The nine topics that emerged from the analysis were breeding, stress and housing conditions, welfare and pain assessment, public health, shelter management and euthanasia, behavioral problems, health issues and management, human-animal interaction, and owners' and veterinarians' perceptions. While stress and housing conditions, public health, and owners' and veterinarians' perceptions were the most studied topics, human-animal interaction was the least studied. This review confirmed the increasing research and interest in cats' and dogs' welfare and showed gaps in knowledge where further studies are needed.