Industrial Relations Dispute Simulation System Prototype with Artificial Intelligence Approach

There are many cases of industrial relations disputes/IRD every year. IRD cases can be resolved by kinship or mediation to reduce the material burden of all parties, such as case costs other than time and energy. With this background, the researcher proposed developing a system for simulating the IRD judicial process with an AI approach. AI encourages faster and high-accuracy prediction results because AI works through the learning process against a set of training data to produce learning models. The research method used is an experimental laboratory to select AI algorithms with the highest accuracy. Meanwhile, for system development methods, research proposes prototyping methods with designing systems using UML. Prototyping is an option because it takes the intensity of communication between the developer and the end user to determine the prototype of the system being built. System development with platform website. UML provides a variety of diagrams that facilitate communication with developers to illustrate the system being built. System testing uses the black box methodology approach because at this stage, testing is carried out to ensure that the functional system has met the needs. From the experimental results, the decision tree algorithm provides the highest accuracy of 80% in training and testing a set of datasets in the form of cases and IRD court rulings from 2022. The accuracy score means that the learning model by the decision tree algorithm can correctly predict (TP / True positive) 75 % of all cases (test data). The accuracy score is obtained through a confusion matrix that shows the performance of the decision tree algorithm for classification. The results of this research help the process of simulating IRD cases before being taken to judicial line to minimize costs and other efforts that could potentially be incurred during judicial process.