topic modeling

Vegetarianism is one of the trends that has received a lot of feedback on social networks. The content published by users reflects their feelings and opinions towards this trend and its various aspects. In this regard, a dataset containing more than sixty thousand tweets published in 2023 about vegetarianism was collected. This dataset was used to extract user sentiment towards different aspects of vegetarianism. First, a method based on RoBERTa language model was proposed to analyze the implicit sentiment hidden in tweets. Then, using the Latent Dirichlet Allocation topic modeling approach, some relevant aspects and topics related to vegetarianism were extracted. In the next step, a method based on DeBERTa language model was used to analyze tweet sentiment towards different aspects that had been extracted. Various frequency and sentiment distribution charts for different aspects in the field of vegetarianism were examined. The results of emotional analysis based on RoBERTa and DeBERTa models were compared side by side. Data analysis using the DeBERTa model showed that users had mostly tweeted positive sentiments regarding the plant and lifestyle aspects. However, for the Animal aspect, most tweets were negative. For both Diet and Company aspects, most tweets were positive or neutral with values close to each other. During the discussion, some implicit knowledge related to this topic was also examined.

Text classification is one of the main tasks of natural language processing (NLP). In this task, documents are classified into pre-defined categories. There is lots of news spreading on the web. A text classifier can categorize news automatically and this facilitates and accelerates access to the news. The first step in text classification is to represent documents in a suitable way that can be distinguishable by a classifier. There is an abundance of methods in the literature for document representation which can be divided into a bag of words model, graph-based methods, word embedding pooling, neural network-based, and topic modeling based methods. Most of these methods only use local word co-occurrences to generate document embeddings. Local word co-occurrences miss the overall view of a document and topical information which can be very useful for classifying news articles.  In this paper, we propose a method that utilizes term-document and document-topic matrix to generate richer representations for documents.  Term-document matrix represents a document in a specific way where each word plays a role in representing a document. The generalization power of this type of representation for text classification and information retrieval is not very well. This matrix is created based on global co-occurrences (in document-level). These types of co-occurrences are more suitable for text classification than local co-occurrences. Document-topic matrix represents a document in an abstract way and the higher level co-occurrences are used to generate this matrix. So this type of representation has a good generalization power for text classification but it is so high-level and misses the rare words as features which can be very useful for text classification. The proposed approach is an unsupervised document-embedding model that utilizes the benefit of both document-topic and term-document matrices to generate a richer representation for documents. This method constructs a tensor with the help of these two matrices and applied tensor factorization to reveal the hidden aspects of data. The proposed method is evaluated on the task of text classification on 20-Newsgroups and R8 datasets which are benchmark datasets in the news classification area. The results show the superiority of the proposed model with respect to baseline methods. The accuracy of text classification is improved by 3%.

With respect to the increasing volume and variety of information available on the Web, it is very difficult to find the required knowledge through the massive amount of data. Question-answering systems have been created to make easy knowledge accessing through massive amounts of data. The most important factor in the issue of expert finding is the ability to detect the relationship between query words and documents written by the candidate experts. A challenging issue in this area is the vocabulary gap between query words and the documents of the candidate experts. In this paper, two new translation models are proposed to solve the problem of the vocabulary gap. First model, a cluster-based probabilistic model, and another is based on topic modeling. In these models, the query words are translated into a collection of query-related words, which are written in documents written by more candidate experts. Then, using these words and using a simple composite model, we have retrieved the experts. The proposed models are implemented and evaluated on the Stack overflow test set and finally, we have analyzed the outputs. The results indicate an increase in the Mean Average Precision of the proposed method compared with other methods of expert finding.