نوع مقاله : مقاله پژوهشی

نویسندگان

• رویا آباده ¹
• علیرضا حسن زاده ²
• شعبان الهی ³

¹ کارشناس ارشد مدیریت فناوری اطلاعات، دانشکده مدیریت و اقتصاد، دانشگاه تربیت مدرس، تهران، ایران

² استاد، گروه مدیریت فناوری اطلاعات، دانشکده مدیریت و اقتصاد، دانشگاه تربیت مدرس، تهران، ایران

³ استاد مدیریت فناوری اطلاعات، گروه مدیریت، دانشکده علوم اداری و اقتصاد، دانشگاه ولی عصر رفسنجان

چکیده

مهمترین اهداف این تحقیق عبارت از ارزیابی و مصورسازی نقشه دانش داده کاوی در تحقیقات کووید-19، شناسایی اجزا و روابط مختلف میان مفاهیم داده کاوی در حوزه های مختلف مطالعات کووید-19 می باشد. این تحقیق از نظر هدف کاربردی و از نوع توصیفی، تحلیلی با بهره مندی از روش تحلیل محتوا و فنون متن کاوی انجام خواهد شد. یکی از یافته های تحقیق حاضر این است که دسته ای از تحقیقات به «ساختار و ماهیت انسانی بیماری» مرتبط بود. دسته ای دیگر از تحقیقات به مجموعه عوامل مربوط به بزرگسالی می پرداخت که می توانست زمینه‌ساز جذب ویروس کووید-19باشد. در بخش هم نویسندگی بیشترین میزان همکاری میان امریکا و هند بوده است. نقشه داده کاوی کووید-19 بر اساس مدلسازی موضوعی مشتمل بر ساختار و ماهیت انسانی بیماری، زمینه و ساختار ویروس، پیشگیری پاندمیک، و علوم کامپیوتری و هوش مصنوعی است. در این تحقیق با توجه به استفاده از ابزارهای فن آوری اطلاعات و تکنیک متن کاوی حجم وسیعی از تحقیقات مورد مطالعه قرار گرفته است. نقشه دانش ارائه شده به تفکیک موضوعی تحقیقات و کشف اطلاعات در هر مجموعه می پردازد لذا امکان مقایسه بین گروهای موضوعی خاص را امکانپذیر می کند. در این تحقیق، اجزای مفاهیم داده‌کاوی در حوزه‌های مختلف مطالعات کووید-19 در شش بعد مورد بررسی قرار گرفت. در بعد ششم نقشه داده کاوی کووید-19 بر اساس مدلسازی موضوعی ارایه گردید.

کلیدواژه‌ها

• نقشه دانش
• داده کاوی
• کووید19
• متن کاوی
• مصورسازی

موضوعات

• علم داده، هوشمندی و تحلیل‌های آینده ‌نگر

عنوان مقاله [English]

Extraction and Visualization of data mining knowledge map in Covid-19 research

نویسندگان [English]

• roya abadeh ¹
• Alireza Hassanzadeh ²
• ُShaban Elahi ³

¹ Senior Expert in Information Technology Management, Faculty of Management and Economics, Tarbiat Modares University, Tehran, Iran

² Professor, Department of Information Technology Management, Faculty of Management and Economics, Tarbiat Modares University, Tehran, Iran

³ Professor in Information Technology Management, Department of Management, Vali-e-Asr University, Imam Khomeini Square, Rafsanjan, Iran.

چکیده [English]

The most important objectives of this research are to evaluate and illustrate the knowledge map of data mining in covid-19 research, to identify the various components and relationships between data mining concepts in different fields of covid-19 studies. The research method is applied and descriptive. In this research, we utilize the content analysis and text mining techniques. One of the findings of the current research is that a group of researches was related to the "structure and human nature of the disease". Another group of research looked at the set of factors related to adulthood that could be the basis for the absorption of the Covid-19 virus. In the co-authoring section, the highest amount of cooperation was between America and India. The data mining map of Covid-19 is based on thematic modeling including the structure and human nature of the disease, the background and structure of the virus, pandemic prevention, and computer science and artificial intelligence. In this research, according to the use of information technology tools and text mining techniques, a large amount of research has been studied. The presented knowledge map deals with the thematic separation of the researches and information discovery in each collection, so it enables the possibility of comparison between specific thematic groups. In this research, the components of data mining concepts in different fields of Covid-19 studies were examined in six dimensions. In the sixth dimension, the data mining map of Covid-19 was presented based on thematic modeling.

Introduction

 What the experience of health officials during the Corona crisis has proven is that it is not possible to manage the disease by relying on traditional methods. Modern information technologies should be used for the optimal management of this disease. Data mining and data analysis can greatly help in preventing, identifying and solving crises. One of the ways that helps researchers to achieve their research goals in their specialized field is to have an overview of the scientific framework of the field in question and previous studies. In this regard, this research has drawn a map of knowledge in the field of Corona data mining by monitoring authors and important and influential works, text mining of studies and specifying thematic clusters formed over time and visualizing information in the field of data mining of Corona. Conducting research in different subject groups provides the possibility of comparing and discovering connections between different fields of knowledge. The provided knowledge map can be used in evidence-based studies in covid-19 researches.

Literature Review

 Considering the large amount of data related to Covid-19 that are produced from various sources such as scientific articles, reports of health organizations, clinical and laboratory data, social networks and media, the need for data mining is felt to extract useful and reliable knowledge (Chen et al., 2020). Data mining can help identify risk factors, diagnose and predict disease, evaluate the effects of treatments and vaccines, analyze social and economic behaviors, and provide solutions (Lee et al., 2020). One of the ways that helps researchers to achieve their research goals in their specialized field is to have a general understanding of the scientific framework of the field in question. In this regard, visualizing information or drawing a map and drawing the scientific structure of that field seems necessary. In a knowledge map that is drawn based on the scientific-research outputs of scientists of a scientific field, influential authors, thematic clusters formed over time and important and influential works are determined and introduced. In knowledge maps, the appearance of new areas and the cessation of some saturated scientific areas can be clearly observed and studied. Anderson (2021) used the text mining clustering approach to evaluate the different foci of a large number of abstracts related to Covid-19. Thakur and Kumar (2021) highlight the applied text mining techniques on published scientific articles by carefully reviewing the studies conducted in the last ten years.

Methodology

This research uses applied methodology in terms of purpose. The statistical population of this research is all the researches published in reliable publications that are in the Scopus database. The statistical sample of this research is all scientific documents registered on the topic of data mining of Covid-19 in the years 2020, 2021 and 2022 in the Scopus database. The number of examined documents is 1749, which includes conference and research articles. In this research, all scientific documents where data mining and covid-19 were simultaneously one of its keywords were extracted. Then, by using appropriate algorithms and, text mining software, knowledge map drawing was done.

Results

 As a result of topic analysis or Topic Modeling with the help of LDA or Latent Dirichlet Allocation algorithm, four topic keyword sets are displayed as described as follows:
 Group 1: includes words that focus on data analysis related to the covid-19 crisis.
 Group 2: includes words that seem to focus on machine learning models used to study covid-19.
 Group 3: includes words about public information and social aspects of the covid-19 crisis.
 Group 4: includes the words also includes data mining and modeling but seems to focus more on the results obtained from these methods in studying different aspects of the crisis.
 In the second part of this study, the co-occurrence map analysis of key words in data mining research on Covid-19 shows the conceptual structure of the relationship between concepts or words in a set of publications.

Discussion

 In order to identify the components and relationships between data mining concepts in different areas of Covid-19 studies, they were identified using text mining and bibliometric methods. As a result of the process of text mining and the application of different algorithms, it can be said that data mining is one of the most widely used techniques in the following 4 areas in the research of Covid-19: 1. Understanding public reactions or crisis impact, 2. Predicting outcomes or analyzing patient data, 3. How to disseminate information through the media and public reactions, and 4. Studying different aspects of the crisis.
 As a result of checking the co-occurrence map of keywords, thematic modeling was divided into the following 4 sub-categories: The first category is in the field of social media, machine learning, and in other words around concepts related to artificial intelligence. Another category is related to topics and concepts such as public health, vaccination, psychology, people's thoughts, communication, fear and emotions. The collection of this cluster can be known as a pandemic. The third category of studies can be considered as "structure and human nature of the disease". The fourth category of studies can be considered related to adulthood. The most widely used data mining method in covid-19 research with the highest reproducibility rate is Classification algorithm with 221 repetitions.

Conclusion

 The components of data mining concepts in different fields of Covid-19 studies were examined in six dimensions. In the sixth dimension, the data mining map of Covid-19 was presented based on thematic modeling.
Keywords: Knowledge Map, Data Mining, Covid-19, Text Mining, Visualization.
 
 

کلیدواژه‌ها [English]

• Knowledge map
• Data mining
• Covid-19
• Text mining
• Visualization