<![CDATA[Catur Insan Cendekia University (UCIC) regularly conducts tracer studies to evaluate graduate outcomes. However, UCIC has not yet implemented an integrated tracer study information system, and alumni data are still managed manually using Microsoft Excel. Consequently, tracer study data, particularly employment waiting period and job field suitability, contain a high proportion of missing values, preventing the university from obtaining a comprehensive graduate profile. This study aims to develop a web-based Tracer Study Information System capable of predicting alumni employment waiting periods, classifying job field suitability for incomplete tracer study records, and analyzing annual trends. The proposed system implements Support Vector Regression (SVR), Support Vector Classification (SVC), and Linear Regression. The dataset consists of 314 tracer study records from the Information Systems Study Program, covering graduation years 2016–2026, and is divided into 270 training records, 28 testing records, and 16 prediction records using a temporal split approach. The SVR model achieved an MAE of 0.8318 months, RMSE of 1.0478 months, and R² of 0.5798, while the SVC model achieved an accuracy of 78.57%, precision of 81.82%, recall of 69.23%, and F1-Score of 75.00%. Linear Regression identified a decreasing trend in both the average employment waiting period (0.2765 month per year) and job field suitability (3.2877% per year). The prediction results were further converted into the Average Graduate Waiting Period (RMT) and Job Field Suitability Percentage (PBS) indicators based on the LAM-INFOKOM 1.0 accreditation instrument. In addition, Black Box Testing confirmed that all system features functioned as expected.]]>
Copyrights © 2026
