Khrystyna Shakhovska

Objectives: This paper introduces an adaptive learning framework for handling concept drift in data by dynamically adjusting model updates based on the severity of detected drift. Methods: The proposed method combines multiple statistical measures to quantify distributional changes between recent and historical data windows. The resulting severity score drives a three-tier adaptation policy: minor drift is ignored, moderate drift triggers incremental model updates, and severe drift initiates full model retraining. Results: This approach balances stability and adaptability, reducing unnecessary computation while preserving model accuracy. The framework is applicable to both single-model and ensemble-based systems, offering a flexible and efficient solution for real-time drift management. Also, different transformation methods were reviewed, and quantile transformation was tested. By applying a quantile transformation, the Kolmogorov–Smirnov (KS) statistic decreased from 0.0559 to 0.0072, demonstrating effective drift adaptation.

An interpretable drift adaptation mechanism for detection and correction is introduced, based on statistical tests and transparent transformations. In contrast to prior work that applies a single universal mapping, the method adaptively selects transformations by drift type (location, scale, shape, or extreme), identified via Kolmogorov–Smirnov tests, Wasserstein distance, and distributional comparisons. Each category is corrected with a suitable transformation such as mean-variance scaling, rank-based adjustment, or quantile mapping. A novel Wasserstein–aware fallback rule ensures balanced corrections across metrics. Applied to salary data across roles and years, the approach reduced Wasserstein distance by over 95% in location+scale drifts (e.g., from 22 416 to 1 118). The method remains easy to interpret, auditable for regulatory checks, and effective for practical drift correction.

Text generation, in particular, next‐word prediction, is convenient for users because it helps to type without errors and faster. Therefore, a personalized text prediction system is a vital analysis topic for all languages, primarily for Ukrainian, because of limited support for the Ukrainian language tools. LSTM and Markov chains and their hybrid were chosen for next‐word prediction. Their sequential nature (current output depends on previous) helps to successfully cope with the next‐word prediction task. The Markov chains presented the fastest and adequate results. The hybrid model presents adequate results but it works slowly. Using the model, user can generate not only one word but also a few or a sentence or several sentences, unlike T9.

The purpose of this paper is to develop a hybrid model Ukrainian language sentiment analyzer, which should improve the accuracy of the mood definition to expand the Ukrainian language among the instruments on the market. The object of research is the processes of determining the language of the text and predicting its sentiment score. The subject of the study is Ukrainian comments posted by Google Maps users. The following text categories are taken into account: food, hotels, museums, and shops. The new method was built as an ensemble of support vector machine, logistic regression, and XGBoost, in combination with a rule-based algorithm. The practical use of the algorithm makes it possible to analyze the Ukrainian text in accordance with the category with the visualization of the research results. The accuracy of the proposed method is bigger than 0.88 in the worst case. The mining procedure of the positive and negative sides of service providers based on users’ feedback is developed. It allows electronics business to make improvements based on frequent positive and negative words.