Multi-record Matching Strategy in Assessment of Complex Information System Skills

Abstract

Complex information system is widely used in modern society, and its skills assessment plays an important role in its practical work. Research on multi-record matching strategy benefits to the quality of assessment results on complex information system skills. This paper takes ERP software as an example and studies the multi-record matching strategy. It firstly designs and implements five record-matching methods, depending on the keywords, the local optimal accuracy, the local optimal scores, the global optimal accuracy and the global optimal scores, and then carries on the comparison to the effect of them. The keywords matching means that all the scoring fields in one record are matched, then the whole record is deemed to be matched. The optimal method refers to selecting the record with the maximum accuracy or scores as the matched record. The local way of matching relies on the order of the records, whereas the global one considers all the possible matches for the determination of the result. When using the accuracy as the optimal selection basis, the number of matching fields in one record must be counted. At the time if the scores are the criteria, the scores of matched fields need to be calculated before working out the final matched records. The paper is aim to help carry out the study in skills assessment and the practical work of complex information system.