Multi-document summarization for human resources performance appraisal

This paper examines machine learning techniques in human resource management, with a focus on 360-degree performance appraisals. We aim to develop salary increase predictive models using logistic regression, Naive Bayes, random forest, and gradient boosting algorithms, utilizing a dataset from Thurein Stores Group that includes employee demographics, work experience, and 360-degree performance appraisal ratings. The data was collected from 176 employees across six branches of Thurein Stores Group through an end-to-end 360-degree performance appraisal process. After data cleaning and preprocessing, sentiment analysis was conducted on the feedback comments from text information across multiple documents. The processed data was then used to train and evaluate the machine learning models. The results show that the proposed gradient boosting model with multi-document remark summarization outperformed other models in terms of precision, accuracy, recall, and F1 score. The proposed gradient boosting-based model achieved nearly 94% accuracy, 94% precision, 94% recall, and 94% F1 score. The study concluded that the proposed feedback text summarization and machine learning models, particularly gradient boosting, can effectively predict salary increases based on performance ratings, providing valuable information for HR decision-making. Future research should investigate the integration of additional variables and the application of these models in various organizational contexts to test their further effectiveness.