The Latent Dirichlet Allocation (LDA) model is a generative model with several applications in natural language processing, text mining, dimension reduction, and bioinformatics. It is a powerful technique in Topic modeling in text mining, which is a data mining method to categorize documents by their Topic. Basic methods for Topic modeling, including TF-IDF, unigram, and mixture of unigrams successfully deployed in modern search engines. Although these methods have some useful benefits, they don’t provide much summarization and reduction. To overcome these shortcomings, the latent semantic analysis (LSA) has been proposed, which uses singular value decomposition (SVD) of word-document matrix to compress big collection of text corpora. User’s search key words can be queried by making a pseudo-document vector. The next improvement step in Topic modeling was probabilistic latent semantic analysis (PLSA), which has a close relation to LSA and matrix decomposition with SVD. By introducing of exchangeability for the words in documents, the Topic modeling has been proceeded beyond PLSA and leads to LDA model. We consider a corpus contains M documents, each document has words, and each word is an indicator from one of vocabularies. We defined a generative model for generation of each document as follows. For each document draw its Topic from and repeatedly for each draw Topic of each word from and draw each word from the probability matrix of with probability of. We can repeat this procedure to generate whole documents of corpus. We want to find corpus related parameters and as well as latent variables and for each document. Unfortunately, the posterior is intractable, and we have to choose an approximation scheme. In this paper we utilize LDA for collection of discrete text corpora. We describe procedures for inference and parameter estimation. Since computing posterior distribution of hidden variables given a document is intractable to compute in general, we use approximate inference algorithm called variational Bayes method. The basic idea of variational Bayes is to consider a family of adjustable lower bound on the posterior, then finds the tightest possible one. To estimate optimal hyper-parameters in the model, we used the empirical Bayes method, as well as a specialized expectation-maximization (EM) algorithm called variational-EM algorithm. The results are reported in document modeling, text classification, and collaborative filtering. The Topic modeling of LDA and PLSA models are compared on a Persian news data set. It has been observed that LDA has perplexity between and, while the PLSA has perplexity between and, which shows domination of LDA over PLSA. The LDA model has also been applied for dimension reduction in a document classification problem, along with the support vector machines (SVM) classification method. Two competitor models are compared, first trained on a low-dimensional representation provided by LDA and the second trained on all documents of corpus, with accuracies and, respectively, this means we lose accuracy but it remains in reasonable range when we use LDA model for dimensionality reduction. Finally, we used the LDA and PLSA methods along with the collaborative filtering for MovieLens 1m data set, and we observed that the predictive-perplexity of LDA changes from to while it changes from to for PLSA, again showing the domination of the LDA method.

The emergence of social media creates opportunities for users to share their thoughts. Billions of short texts are produced on social media daily, and their analysis in text mining and content analysis is essential. Detecting Topics from short texts on a large scale is an important and challenging task. Few studies have been conducted on Topic detection in Persian short texts, and the existing algorithms are not remarkable. Therefore, we decided to study the Topic detection in Persian. Topic modeling is a Topic detection technique that extracts groups of words as Topics from documents. Recently, neural Topic models have shown improvements in increasing the coherence of Topic modeling. Also, text embeddings have enhanced neural models. For this reason, in this study, two combined Topic models and the ZeroShot Topic model are presented for Topic detection in Persian social media short texts. These two models incorporate pre-trained BERT text representation into neural Topic models. The experimental results show that these two methods outperform the comparison methods with the highest F1-measure, Topic diversity, and coherence score. Also, the ZeroShot Topic model has better results in terms of evaluation metrics than the combined Topic model

Objective: According to the variety and diversity of research conducted in the field of internal auditing, the purpose of this research is to gain a deeper understanding of the Topic by conducting a systematic review of existing studies to classify Topics and outline some potential future research directions. Methods: In order to achieve our purpose and to more efficiently examine a large number of articles, the "Computational Literature Review (CLR)" approach is used, and through this, the content of abstracts is analyzed automatically to provide a set of research Topics within internal audit literature (Topic modeling). The required information is collected from the Thomson Reuters Web of Science (WoS) database for 1401 articles over almost one hundred years (1920 to 2021). Results: The results of Topic modeling show the focus of seven areas in internal audit research as follows: internal audit in the healthcare industry; internal audit in the areas of risk management, fraud, and internal controls; competence and training and internal audit quality; and the relationship between the external auditor and the internal auditor. A separate review of each of these identified issues related to internal audit reveals the most critical research conducted in this area and identifies some avenues for future research. Conclusion: The results of this study indicate that throughout the evolution of internal auditing during its history (almost 100 years), its philosophy has evolved from ensuring the accuracy of financial statements and accounts and preventing the misuse of corporate funds or assets by helping to create and add value to companies. Internal auditing services have been upgraded from assuring financial statements and responding to its limited stakeholders (i. e., senior managers), giving assurance, consulting, forecasting risk, and being accountable to all stakeholders. Also, the change in the accountability direction of internal audit and its greater cooperation with external auditors, and the growth of technology and increasing the complexity of organizations have led to changes in competencies and skills required and improving and promoting its independence and objectivity; also engaging internal auditors in other areas of expertise (other than accounting and auditing), then it is expected that the trend will continue in the future.

The purpose of this study is to explore the thematic trend analysis of Iranian articles in Library and Information Science based on Topic modeling (LDA) and linear regression model. The population of this study consists of 709 articles indexed in Scopus during 2008-2009. In order to achieve the research objectives, the data were analyzed using text mining algorithms, especially LDA thematic modeling algorithms using R software. The results showed that among the extracted Topics, there are Topics that have high research popularity and are considered as hot Topics. These Topics include library services on social media, research models, social capital, medical databases, data mining, scientific production trends, interdisciplinary studies, cyberspace algorithms, knowledge management, social media studies, research approaches, and future studies. Also, Topics that are less popular and are considered as cold Topics include areas such as electronic resources, information management system, search engines, book loan services, distance services, e-learning, e-government, journal evaluation indicators, evaluation of web resources, and digital libraries. The results indicated that Library and Information Science research in Iran has developed in line with the growth of technologies and global Topics and has established the relationship between Library and Information Science subject area and new fields of data mining, artificial intelligence, semantic retrieval, ontologies, information architecture, digital publishing, social networks, and databases.

Sentiment analysis is a process through which the beliefs, sentiments, allusions, behaviors, and tendencies in a written language are analyzed using Natural Language Processing (NLP) techniques. This process essentially comprises of discovering and understanding people's positive or negative sentiments regarding a product or entity in the text. The increased significance of sentiments analysis has coincided with the growth in social media such as surveys, blogs, Twitter, etc. The present study takes advantage of the Topic modeling approach based on latent Dirichlet allocation (LDA) to extract and represent the thematic features as well as a support vector machine (SVM) to classify and analyze sentiments at the aspect level. LDA seeks to extract latent Topics by observing all the texts, which is accomplished by assigning the probability of each word being attributed to each Topic. The important features that represent the thematic aspect of the text are extracted and fed to a support vector machine for classification through this approach. SVM is an extremely powerful classification algorithm that provides the possibility to separate complex data from one another accurately by mapping the data to a space with much larger aspects and creating an optimal hyperplane. Empirical data on real datasets indicate that the proposed model is promising and performs better compared to the baseline methods in terms of precision (with 89. 78% on average), recall (with 78. 92% on average), and F-measure (with 83. 50% on average)

ObjectiveThe rapid growth and complexity of research at the intersection of accounting and sustainability necessitate a structured approach to synthesizing existing knowledge. Traditional literature reviews, while valuable, are limited in comprehensively mapping this multidisciplinary field and capturing its evolving thematic structures. This study employs Latent Dirichlet Allocation (LDA) Topic modeling to systematically map the intellectual landscape of research at the intersection of accounting and sustainability. It identifies dominant and emerging themes, traces their interconnections and temporal dynamics, and highlights key research clusters and knowledge gaps. The findings provide evidence-based guidance for researchers, propose avenues for future inquiry, and contribute to shaping a more coherent research agenda, while offering practical insights for professionals, managers, and policymakers. MethodsThis research employs a quantitative, computational approach that integrates bibliometric analysis with advanced Topic modeling. Data were collected through a systematic search of the Scopus database using comprehensive accounting-and sustainability-related keywords, covering the period from October 1969 to May 2025. The initial search retrieved 9, 863 documents. To ensure analytical precision, a multi-stage preprocessing pipeline was applied to abstracts, including tokenization, punctuation removal, lowercasing, and a two-tiered stop-word removal procedure (combining a standard English list with a custom list of common academic terms). After cleaning, the final dataset comprised 8, 913 articles, which were modeled using the LDA algorithm. The optimal number of Topics (k = 12) was determined based on the highest coherence score (C_v metric). All analyses were conducted in Python using the Gensim and NLTK libraries. ResultsBibliometric analysis indicates sustained growth in scholarly output, with publication volumes peaking in 2023 and 2024, reflecting heightened academic engagement. Citation peaks in 2009 coincided with the global financial crisis and the emergence of carbon accounting, while 2021 was shaped by the COVID-19 pandemic and accelerated transitions toward a low-carbon economy. LDA Topic modeling reveals a rich and interconnected landscape of research, integrating environmental, social, governance, and reporting dimensions. Studies on life cycle assessment and environmental impacts of production connect closely with carbon emissions and climate change mitigation, reflecting efforts to quantify and address the environmental footprint of business operations. Sustainability frameworks and social dimensions intertwine with corporate social responsibility (CSR) reporting practices and sustainability reporting and ESG disclosure standards, highlighting the interplay between normative principles and corporate disclosure mechanisms. Meanwhile, research on green supply chains and sustainability challenges in emerging economies underscores sectoral and geographic variations in implementing sustainability practices, bridging operational and contextual insights. Corporate-level investigations, including ESG and corporate financial performance, corporate governance and board composition, and assurance and sustainability in financial institutions, reveal how governance structures and performance metrics intersect with sustainability objectives. At the policy and public sector level, studies on green policies and public sector sustainability illustrate regulatory and institutional influences on sustainable practices. Finally, research on sustainability in business education and tourism signals a growing attention to shaping future professionals’ competencies and extending sustainability considerations beyond traditional corporate contexts. Together, these interconnected themes reveal an evolving field in which conceptual foundations, environmental imperatives, corporate reporting mechanisms, governance structures, and educational initiatives collectively define the trajectory of accounting and sustainability research. ConclusionThe analysis demonstrates a complex, dynamic, and integrated intellectual landscape, emphasizing the necessity of embedding Environmental, Social, and Governance (ESG) considerations into core business practices. This integration redefines concepts of value, accountability, and long-term viability, while presenting challenges for professional practice and regulatory frameworks. Current research momentum—particularly in ESG and corporate financial performance and sustainability reporting, and ESG disclosure standards—reflects a strong focus on the financial and communicative dimensions of sustainability. Environmental themes, including carbon emissions and life cycle assessment, highlight continued engagement with quantifying impacts, supporting reliable accounting and decision-making. The field’s scope extends beyond corporate boundaries to include sustainable supply chains, emerging economies, public sector sustainability, and the integration of sustainability into business education. Overall, research shows a progression from broad conceptual frameworks toward more granular, sector-specific, and assurance-focused studies, increasingly shaped by regulatory influence, with all themes functioning as an interconnected whole that advances understanding and practice in accounting and sustainability.

This article presents a method to find the number of Topics in Persian research articles, which is actually one of the main challenges in Topic modeling. It is the process of automatically recognizing Topics in a text with the aim of discovering hidden patterns. This study has estimated the number of Topics for Persian research articles using two approaches. The first is based on the greedy search and later uses Renormalization theory, which is a mathematical formalism to construct a procedure for changing the scale of the system so that the behavior of the system preserves. Also, the execution time of both algorithms on Persian academic articles has been compared with each other. The findings indicate that the renormalization approach predicts the number of Topics in Persian research articles with the lower time complexity in comparison to the greedy based approach. The approach based on Renormalization has high efficiency for estimating the number of Topics in Persian academic articles.