<![CDATA[Financial distress is a critical phenomenon in a company that has significant implications for the business itself, employees, investors, and creditors, and can also impact the economy of a country. Predicting the financial distress of a company, including property and real estate companies, becomes one of the crucial things to be studied. The Support Vector Machine (SVM) is said to be the most effective model for prediction and classification among other machine learning methods. However, it is difficult to determine the parameters of the SVM model. Thus, the SVM model's parameters must be improved for higher accuracy results. This research aims to increase the accuracy of the SVM model in predicting the financial distress of property and real estate companies. The optimization method used is Particle Swarm Optimization (PSO). PSO is one of the most well-known techniques for enhancing SVM parameters. The PSO approach takes its cues from how a group of insects or birds interacts to maintain life. Initialized in a D-dimensional search space, the PSO algorithm uses a population of random particles that are considered as points. Each particle modifies its direction using the best experience it discovers (pbest) and the best experience discovered by all other members (gbest) to arrive at the ideal outcome. As a result, throughout the search process, particles will move through multidimensional space to more advantageous locations. The result of this research showed that the SVM model has the highest accuracy at 80.47% while when the PSO method was implemented in the SVM model, the accuracy increased into 83.16%. It can be concluded that the PSO method successfully optimized the parameters and increased the accuracy of SVM model in predicting the financial distress of property and real estate companies listed in Indonesian Stock Exchange.]]>
Copyrights © 2024
