C4.5 Classification Algorithm Based On Particle Swarm Optimization To Determine The Delay Order Production Pattern

Abstract

Manufacturing companies engaged in the production of cosmetic products and herbal products on a national and international scale, this company is required to be able to carry out production planning by using classification methods in determining the pattern of delay in production orders for the C4.5 classification algorithm based on Particle Swarm Optimization. The research method used in this experiment is the Cross Industry Standard Model for Data Mining (CRISP-DM). From the results of the C4.5 algorithm optimization analysis with particle swarm optimization (PSO), it can be concluded that the accuracy value obtained from the PSO-based C4.5 algorithm model is 82.52%. This is better than the C4.5 algorithm model which produces an accuracy value of 80.17%. The Difference of 0.033 with the details of the C4.5. Algorithm producing an AUC value of 0.855 with a diagnosis of Good Classification. The results of the optimization analysis of the C4.5 algorithm with PSO, the pattern of delay that is formed is that the production status is the status that causes the most delay in production orders. then followed by the status of waiting for pre-production and administrative status. the emerging status must be a correction and attention so that management becomes better.

Study of The ID3 and C4.5 Learning Algorithms

“…In the next section, we will discuss in depth the measures behind the constructing of the model, then later we will present how to make these measures work with learning algorithms ID3 [1,2] and C4.5 [3,4,5,6].…”