An intelligent module for productivity assessment of remote employees in working time monitoring systems

This paper presents the development of an intelligent module for a working time management and accounting system aimed at improving remote employee productivity analysis and decision support. The proposed module extends traditional time tracking functionality by integrating multicriteria productivity assessment, anomaly detection, short-term forecasting, and recommendation generation mechanisms. The developed approach combines an AHP-based productivity evaluation model, a hybrid anomaly detection algorithm using the Isolation Forest and statistical analysis, an LSTM-based forecasting model, and a fuzzy logic recommendation system. The developed methods were verified using a synthetic dataset containing 500 records. The obtained results confirmed the effectiveness of the proposed solution. The productivity assessment model demonstrated a strong correlation with expert evaluations (r = 0.87), while the anomaly detection and forecasting components showed satisfactory performance according to the selected evaluation metrics. The recommendation subsystem demonstrated practical applicability, with 78% of generated recommendations accepted by management personnel. The developed intelligent module provides an integrated approach to productivity analysis, detection of abnormal work patterns, forecasting of future performance, and support of managerial decision-making in remote workforce management systems.