Please use this identifier to cite or link to this item:
|Title||Detecting Subjectivity in Staff Performance Appraisals Using Text Mining (Teachers’ Appraisals of Palestinian Government Case Study)|
|Title in Arabic||اكتشاف عدم الموضوعية في تقييمات اداء الموظفين باستخدام التنقيب عن النصوص|
As human resources are the resources that carry out many important activities in an organization, Human Resource Management (HRM) should catch up with the latest developments to manage these resources efficiently. Staff appraising is one of the most important roles of HRM. Accurate appraising systems promise organizations of a plethora of benefits. Right managerial decisions and staff's perception of fairness are some of these benefits. Non-subjective appraising is such a characteristic of accurate appraising systems. However, almost already applied processes for ensuring non-subjectivity in staff appraisals are manual, infeasible, hard and time consuming. For large organizations with large number of staff such as the Palestinian government, it become more difficult. A considerable effort has been directed to detecting subjectivity in opinion reviews. However, to the best of our knowledge, there is no previous work that detect subjectivity in staff appraisals. Our contribution in this work is to use text mining methods in finding context and domain driven clues of subjectivity in staff appraisals. The objective of this work is to propose a text mining based approach that supports HRM in detecting subjectivity in staff performance appraisals. The approach detects three clues of subjectivity in reviews, where each clue represents a level of subjectivity. First level, textual reviews that are irrelevant to the domain of staff appraising. Second level, duplication and near duplication in reviews. Third level, textual reviews that do not provide significance meaning; nearly a duplication of items in the non-textual part of appraisal. For proving our approach, we applied it on the teachers’ staff appraisals of the Palestinian government. According to our experiments, we found that the approach is effective regarding our evaluations; where we used expert opinion, precision, recall, accuracy and F-measure. In the first level, we reached the F-measure of 88%, in the second level, we used expert staff’s opinion, where they decided the percent of duplication to be 85% and in the third level, we achieved the best average F-measure of 84%.
|Publisher||الجامعة الإسلامية - غزة|
|Files in this item|