topic modeling

IT's rapid progress and far-reaching impact on other scientific disciplines have not only necessitated significant changes in its own subjects but have also catalyzed extensive changes in the form, amount, and methodology of research in other fields. The objective of the present investigation was to analyze research conducted in the realm of information technology, extract its central themes, and furnish scientometric data pertaining to these themes.

This paper explores the topics of the information technology field by extracting and categorizing the relationships between commonly used terms and their temporal evolution. To achieve this, the researchers employed topic modeling, a well-established method for clustering textual data. Topic modeling algorithms utilize statistical methods to analyze and interpret the primary words in documents, allowing for the examination of the presented issues and their interconnections and changes over time. Considering the rapid changes in the field of information technology, this paper drew upon materials spanning the last decade, including 10,000 papers sourced from top-tier journals featured in the Web of Science database.

The study extracted trends in keyword changes over the past decade and identified important keywords for each paper cluster after grouping them. The papers within the information technology domain were then categorized into eight themes, including hardware, communications, networks, and intelligent applications such as the Internet of Things. The study found that frequently used keywords have been continuously changing over time. The paper highlights that emerging keywords, including the Internet of Things, cloud computing, and Big Data, along with work areas such as Machine Learning and Deep Learning, are shaping the definition of information technology fields in the new era.

Given the shift in research emphasis from hardware and communication to analysis and practical applications, it is likely that future scientific fields will focus on creating value through data analysis and human-machine communication in everyday applications, and information technology's relevance in other sciences will become more apparent. Future research can also concentrate on comparing global trends in information technology with domestic research, enabling the evaluation of the gap between the country's research and that of the world.

This study investigates the automatic keyword extraction from the table of contents of Persian e-books in the field of science using LDA topic modeling, evaluating their similarity with the golden standard, and users' viewpoints of the model keywords.

This is mixed text-mining research in which LDA topic modeling is used to extract keywords from the table of contents of scientific e-books. The evaluation of the used approach has been done by two methods of cosine similarity computing and qualitative evaluation by users.

Table of contents are medium-length texts with a trimmed mean of 260.02 words, about 20% of which are stop-words. The cosine similarity between the golden standard keywords and the output keywords is 0.0932 thus very low. The full agreement of users showed that the extracted keywords with the LDA topic model represent the subject field of the whole corpus, but the golden standard keywords, the keywords extracted using the LDA topic model in sub-domains of the corpus, and the keywords extracted from the whole corpus were respectively successful in subject describing of each document.

The keywords extracted using the LDA topic model can be used in unspecified and unknown collections to extract hidden thematic content of the whole collection, but not to accurately relate each topic to each document in large and heterogeneous themes. In collections of texts in one subject field, such as mathematics or physics, etc., with less diversity and more uniformity in terms of the words used in them, more coherent and relevant keywords are obtained, but in these cases, the control of the relevance of keywords to each document is required. In formal subject analysis procedures and processes of individual documents, this approach can be used as a keyword suggestion system for indexing and analytical workforce.

Probabilistic topic modeling methods consist of a set of algorithms whose main purpose is to discover the hidden subject structure in a large volume of documents. The purpose of this study is to thematically model the articles of Iranian researchers in the field of endocrinology and metabolism in the citation database of Web of Science.

The present research is of applied type and has been done by text mining and content analysis method. In this study, all required data were retrieved from the Web of Science Citation Database using the keywords registered in the medical subject heading without a time limit until November 6, 2018. Then, using a hidden allocation algorithm, the whole set of documents in MATLAB was analyzed.  

Subject categories were extracted as groups of 20 words in 10 subject categories. Then, by endocrinologists, the subject categories were named based on their relationship to various topics in the field of  endocrinology and metabolism, and each category was assigned a subject title.

The results indicate that the implementation of the latent Dirichlet allocation model has an acceptable performance in presenting the categories of endocrinology and metabolism. The extracted subject categories have good homogeneity and thematic relevance with each  other.

Topic modeling is one of the text mining techniques that allows you to discover unknown topics in a collection of documents, interpret documents based on these topics, and use these interpretations to organize, summarize, and search for texts automatically. Familiarity with the concept and technique of topic modeling, and its application in discovering topics and organizing information is one of the main goals of this research.

The present study is a review-analytical type in which, while introducing topic modeling, it has categorized and reviewed the applications of this technique based on its performance and provided a sample of research that has used this technique.

Topic modeling algorithms is used not only in addition to the three main objectives of discovering hidden topics, interpreting documents based on topics, and finally organizing and classifying texts, but also is used in discovering hidden topics and relationships in the fields of science, information retrieval, categorizing documents based on topics, discovering outstanding patterns and emerging events, clustering the concepts of scientific fields, analyzing the course of conceptual evolution during historical periods, determining the hierarchical relationships of concepts. A specific scientific field or field and vocabulary enrichment.

Topic modeling based on machine learning and artificial intelligence knowledge has been proposed as one of the new approaches to organizing information resources and serious studies are being conducted in this field. Therefore, by using topic modeling algorithms in order to automate the extraction of the subject and discover the hidden issues in the source, it is possible to strengthen and update the new systems of organizing information resources.

The purpose of this study is to explore the subject structure and hidden knowledge and development of new research subjects. This research is applied from the point of purpose and type, descriptive and from the point of method, trend analysis. The population of this study is the all of articles indexed in Scopus and written by Iranian authors in the field of knowledge and information science from 1992-2019, Then take the targeted sample in 2008-2019.Data analysis tool is software R which abstract the papers were analyzed after applying pre - processing steps with Topic modeling and LDA algorithm. The result of study to identification topic development process showed Hot topics & Cold topics to identifying the process of knowledge and information science by determining the p-level (0/05) and identify alpha, beta and theta parameters. The Hot topics that have a more popularity is: Library services on Social Networks, Research Models, Social Capital, Data Base, Data Mining, Trend of Scientific Publication, Interdisciplinary Fields, Cyberspace Algorithms, Knowledge Management and Cold topic that have a less popularity is: Electronic Resources, Information management System, Search Engine, Book loan Services, Distance Library Services, E learning, E Government, Journal Evaluation Indicators, Web sources Evaluation, Digital Library. The result of this research showed that the field of knowledge and information science can growth up with globalization and new global issues and this field ability to develop and expand in data mining, artificial intelligence, semantic retrieval, information architecture, digital publishing, social networking and information technology.

Four text mining methods are examined and focused on understanding and identifying their properties and limitations in subject discovery.

The study is an analytical review of the literature of text mining and topic modeling. 

LSA could be used to classify specific and unique topics in documents that address only a single topic. The other three text mining methods focus on topics and general partiality of the text. PLSA is applicable to documents dealing with a topic, unlike the LSA, it is used to discover general themes and contexts. However, LDA is more applicable to documents that address several issues. The CTM, method can be used to identify relationship between different subject categories.

Text mining tactics are suitable for employing analysis in discovering and extracting the text subjects.

Improvement in information retrieval performance relates to the method of knowledge extraction from large amounts of text information on web. Text classification is one of application of knowledge extraction with supervised machine learning methods. This paper proposed Kullback-Leibler divergence KNN for classifying extracted features based on term weighting with Latent Dirichlet Allocation Algorithm. LDA is Non Negative matrix factorization method proposed for topic modelling and dimension reduction of high dimensional feature space .In traditional LDA, each component value is assigned using the information retrieval TF measure, While this weighting method seems very appropriate for IR, it is not clear that it is the best choice for TC problems. Actually, this weighting method does not leverage the information implicitly contained in the categorization task to represent documents. In this paper, we introduce a new weighting method based on Point wise Mutual Information for accessing the importance of a word for a specific latent concept, then each document classified based on probability distribution over the latent topics. Experimental result investigated when we used PMI measure for term Weighing and  KNN with Kullback-Leibler  distance, accuracy has been 82.5%, with lower complexity and same accuracy versus complex deep learning methods.