Automatic Topic detection seems unavoidable in social media analysis due to big text data which their users generate. Clustering-based methods are one of the most important and up-to-date categories in Topic detection. The goal of this research is to have a wide study on this category. Therefore, this paper aims to study the main components of clustering-based-Topic-detection, which are embedding methods, distance metrics, and clustering algorithms. Transfer learning and consequently pretrained language models and word embeddings have been considered in recent years. Regarding the importance of embedding methods, the efficiency of five new embedding methods, from earlier to recent ones, are compared in this paper. To conduct our study, two commonly used distance metrics, in addition to five important clustering algorithms in the field of Topic detection, are implemented by the authors. As COVID-19 has turned into a hot trending Topic on social networks in recent years, a dataset including one-month tweets collected with COVID-19-related hashtags is used for this study. More than 7500 experiments are performed to determine tunable parameters. Then all combinations of embedding methods, distance metrics and clustering algorithms (50 combinations) are evaluated using Silhouette metric. Results show that T5 strongly outperforms other embedding methods, cosine distance is weakly better than other distance metrics, and DBSCAN is superior to other clustering algorithms.

Metalepsis, in its narratological sense, is a trope in which an unnatural relationship is built between different levels of narrative. The natural relationship between narrative levels is formed by the act of narrating; a character from one level becomes the narrator of another. The term “ Metalepsis of Topic and illustration” can be coined to name a similar trope. This trope has been used for centuries in Persian poetry. Every Image comprises a Topic and an illustration. The Topic is what is being talked about and the illustration is what the Topic is compared to. When several images are present along together, two different levels are distinguishable: the level of the Topic and the level of illustration. The natural relationship between these two levels is similarity and any other relationship will result in metalepsis. As in the narratological metalepsis, there is always a paradox in the metalepsis of Topic and illustration. The effect of these two kinds of metalepsis is also similar and can be humorous, fantastic, or a mixture of the two.

This paper investigates text documents regarding their Topic density. It has divided them into two groups: dense and sparse documents. Dense documents are texts with a wide domain of Topics. They have a high Topic density (for example religious books, encyclopedia, magazine archives, etc ). We have shown that a) traditional methods cannot be used for Topic specific of dense texts, and b) we can benefit from employing the efficiency of the proposed method (Nasir) for dense texts.In this research, we have used dependency relations, paths, triple databases and statistical text processing methods to extract important words and to insert them into a clustering index. Also a method was described to find the reference of pronouns in dense texts. In addition, based on the suggested methods, a prototype system called Nasir was implemented. The result of the implementation on Persian dense texts shows that the quality of indexing and searching improved significantly.

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.

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 purpose of this research is to compare the prosodic structure of the different types of Persian Topic constructions within the framework of the auto-segmental metrical phonology (Ladd, 2008) in order to answer the challenge of whether the initial position in these constructions is the result of base-generation or the result of the movement of the initial element. In this regard, following Cinque(1990), we examine the Persian Topic constructions in the three categories of Topicalization,  clitic left dislocation (CLLD) and hanging Topic left dislocation (HTLD) in order to determine whether the initial element in these constructions forms an independent intonational phrase or is a part of a larger intonational phrase involving the whole sentence. For this purpose, we compare the values of the phonetic parameter of initial element final syllable length and the amount of pause between this element and the continuation of the clause in different types of Topic constructions. according to Pierrehumbert (1980: 20), if the lengthening of the final syllable of the target group and the duration of the pause between it and the next words are more than these values in the unmarked construction, that group forms an independent intonational phrase. The results show that in Persian Topicalization and CLLD, these values are not significantly different from such values in the unmarked construction; Also, like the unmarked construction, the initial element in Topicalization and CLLD is produced with H- boundary tone. These features can be seen as the phonetic correlation of the strong syntactic connection between the initial element and the continuation of the clause in these two constructions; A characteristic that indicates the movement of the element to the beginning of the clause. On the other hand, the more lengthening of the final syllable of the initial element and the greater amount of pause between this element and the continuation of the clause in HTLD compared to the unmarked construction, is evidence that the initial structure in this construction forms an independent intonational phrase; A characteristic that, along with L- boundary tone, is the phonetic representation of weak syntactic connection in this construction. This feature strengthens the assumption that HTLD is the result of the base-generation of the initial element. The results of this research show that the characteristics of Topic constructions that affect the subject of sentence follow the same pattern of the characteristics of Topic constructions from the object of sentence.