<![CDATA[Identifying interests, talents, and determining career plans is a common challenge faced by adolescents, especially at the upper secondary education level. This study aims to develop a mobile application to identify interests, talents, and career planning using the forward chaining method to analyze user responses. The system’s knowledge base was constructed by adapting Holland’s RIASEC theory, refined through interviews with psychology experts. Instrument validation used the percentage-of-agreement parameter, which, after revisions based on expert feedback, achieved a 100% agreement rate. These validation results were transformed into 38 IF–THEN inference rules that map 30 questionnaire items to six interest categories. The application was developed as a cross-platform mobile system integrated with cloud services for authentication and data storage. The forward chaining inference engine operates through four stages: receiving user response facts, matching patterns against the rule base, executing satisfied rules, and drawing conclusions by calculating the accumulation of rules in each category to determine three dominant categories as the identification result. The prototype implementation includes a job-trend feature based on data from the Ministry of Manpower (Kemnaker) for 2022–2024 as additional support for interest and talent identification. Functional testing showed a 100% success rate across all features. System accuracy evaluation using 20 test data points and a confusion matrix produced an Accuracy of 61.9%, Precision of 23.1%, and Recall of 33.3%. The dominance of True Negative values (23 cases) indicates that the system’s main strength lies in its filtering capability to eliminate irrelevant career options. The low precision value reflects the system’s inclusive design, intended to detect multipotentiality that may not yet be recognized through users’ subjective perceptions. These findings indicate that the application of the forward chaining method in a mobile-based system has the potential to assist users in recognizing their interests and talents more systematically; however, the limitations of binary logic in capturing gradients of interest intensity suggest the need for developing a hybrid model by integrating uncertainty-based methods such as the Certainty Factor or Fuzzy Logic to improve diagnostic sensitivity in future research.]]>
