2023 № 1 Системний аналіз, управління та інформаційні технології
Постійне посилання колекціїhttps://repository.kpi.kharkov.ua/handle/KhPI-Press/67240
Переглянути
2 результатів
Результати пошуку
Документ Using long short-term memory networks for natural language processing(Національний технічний університет "Харківський політехнічний інститут", 2023) Onyshchenko, Kostiantyn; Daniiel, YanaThe problem of emotion classification is a complex and non-trivial task of language interpretation due to the natural language structure and its dynamic nature. The significance of the study is in covering the important issue of automatic processing of client feedbacks, collecting opinions and trendcatching. In this work, a number of existing solutions for emotion classification problem were considered, having their shortcomings and advantages illustrated. The evaluation of performance of the considered models was conducted on emotion classification on four emotion classes, namely Happy, Sad, Angry and Others. The model for emotion classification in three-sentence conversations was proposed in this work. The model is based on smileys and word embeddings with domain specificity in state of art conversations on the Internet. The importance of taking into account the information extracted from smileys as an additional data source of emotional coloring is investigated. The model performance is evaluated and compared with language processing model BERT (Bidirectional Encoder Representations from Transformers). The proposed model achieved better performance at classifying emotions comparing to BERT (having F1 score as 78 versus 75). It should be noted, that further study should be performed to enhance the processing by the model of mixed reviews represented by emotion class Others. However, modern performance of models for language representation and understanding did not achieve the human performance. There is a variety of factors to consider when choosing the word embeddings and training methods to design the model architecture.Публікація An algorithm for NLP-based similarity measurement of activity labels in a database of business process models(Національний технічний університет "Харківський політехнічний інститут", 2023) Kopp, Andrii Mykhailovych; Orlovskyi, Dmytro LeonidovychBusiness process modeling is an important part of organizational management since it enables companies to obtain insights into their operational workflows and find opportunities for development. However, evaluating and quantifying the similarity of multiple business process models can be difficult because these models frequently differ greatly in terms of structure and nomenclature. This study offers an approach that uses natural language processing techniques to evaluate the similarity of business process models in order to address this issue. The algorithm uses the activity labels given in the business process models as input to produce textual descriptions of the associated business processes. The algorithm includes various preprocessing stages to guarantee that the textual descriptions are correct and consistent. First, single words are retrieved and transformed to lower case from the resulting textual descriptions. After that, all non-alphabetic and stop words are removed from the retrieved words. The remaining words are then stemmed, which includes reducing them to their base form. The algorithm evaluates the similarity of distinct business process models using similarity measures, including Jaccard, Sorensen – Dice, overlap, and simple matching coefficients, after the textual descriptions have been prepared and preprocessed. These metrics provide a more detailed understanding of the similarities and differences across various business process models, which can then be used to influence decision-making and business process improvement initiatives. The software implementation of the proposed algorithm demonstrates its usage for similarity measurement in a database of business process models. Experiments show that the developed algorithm is 31% faster than a search based on the SQL LIKE clause and allows finding 18% more similar models in the business process model database.