Browsing by Author "Iskalinov F."

Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    A PROFESSION RECOMMENDER SYSTEM BASED ON DEEP LEARNING AND MACHINE LEARNING APPROACHES
    (СДУ хабаршысы - 2023, 2023) Iskalinov F.
    Abstract. The issue of uncertain career path choice among modern schoolchildren has become increasingly prominent, resulting in a substantial decrease in the number of university students. This uncertainty has become a major concern as students and their parents are often unfamiliar with the wide range of available professions, particularly those that have emerged in the last decade. A modern solution is proposed in the form of a web application that uses Deep Learning, Machine Learning, and NLP to recommend suitable specialties based on the competencies required for the profession. The system will analyze and extract implicit features through a supervised classification approach, providing a comprehensive solution for profession search in the Kazakhstan market.
  • Thumbnail Image
    ItemOpen Access
    Implementing automatic word prediction and autocorrection for the Kazakh-language keyboard on the iOS platform
    (2023) Iskalinov F.
    In the past few years, there has been a growing interest in natural language processing (NLP) and the development of language models for various applications, including text prediction and word correction in mobile device keyboards, aimed at improving user experience. However, iOS platforms lack special features adapted for Kazakh-speaking users. The main goal of this thesis is to develop and evaluate word prediction and autocorrection systems for the Kazakh language. Advanced methods such as LSTM and GRU are utilized to efficiently gather contextual information and predict the next word. Various error correction methods, including edit distance, N-gram based models, and a hybrid approach, are applied and evaluated. Experimental analysis using LSTM and GRU models allows for the optimization of hyperparameters and the improvement of word prediction accuracy. It was observed that the LSTM model achieved the highest accuracy. The hybrid approach achieves the highest accuracy of 91% among the evaluated error correction methods. Moreover, the integration of these models and methods with the iOS system enables the development of a fully-featured keyboard application specifically designed for the Kazakh language. This research contributes to the advancement of predictive and auto corrective technologies for next word prediction in Kazakh language processing, aiming to enhance the accessibility and usability of Kazakh language applications. It ultimately improves the overall quality of written communication and enhances user satisfaction with mobile keyboard applications.