Analysis of Semantic Relationships in Ukrainian Text Content Based on Word2Vec and Machine Learning

The study used a corpus of Ukrainian texts from books by different authors to train the Word2Vec model to identify semantic relationships between words. The project aims to analyse semantic relations between words and to construct vector representations of words using the word2vec method. The tasks of the research include data preparation, training of the model, analysis of the obtained vectors and comparison of the results with existing methods. Large amounts of high-quality Ukrainian language datasets for training the model, creating an effective word2vec learning algorithm and developing the user interface are the main challenges of the project. Using Python and TensorFlow, an effective application of the word2vec method is achieved, showing high performance. It runs for about 2.8 seconds and uses a limited amount of memory, ranging from 140 to 200 megabytes.