THE DEVELOPMENT URGENT DECISION MODEL FOR READY MIXED CONCRET TRUCK ACCIDENT BY USING MACHINE LEARNING TECHNIQUES

Abstract

        The objectives of this research were to develop urgent decision model for ready mixed concrete truck accident, to test the efficiency of urgent decision model for ready mixed concrete truck accident, and to evaluate urgent decision model for ready mixed concrete truck accident. The data used in this study was collected from the information about accidents involving ready mixed concrete transport truck in one location of incidents of establishments in the Khlong Sam Wa Area in Bangkok, which included incidents from 2019 to 2020 in total of 1,000 items. Evaluation of the urgent decision model for ready mixed concrete truck accident was performed by using the sample of establishments in the Minburi and Khlong Sam Wa Areas in Bangkok with purposive selection to obtain 40 establishments that met the criteria. The research tools used were interview forms, questionnaires, and machine learning techniques. The results from synthesizing the factors were used to create a model for urgent decision model in ready mixed concrete truck accident using machine learning. The results revealed that the UDM Model was developed, and the model consisted of problem identification, severity assessment, information notification, process, solution, action and assessment. Efficiency testing of urgent decision model in ready mixed concrete truck accident using machine learning techniques with an Artificial Neuron Network split learning method into training for 70 percent and test for 30 percent with the highest accuracy of 89.52 %. The learning ability of the system users who were able to use the system well by themselves was at a highest level of 75%. As for the ability to recognize the use of the system users, the users could easily remember the form and method of using the system at a high level of 55%. As for system performance, it could help solve users’ problems at a highest level of 75%. As for system errors, there were no errors found, and it was found at a highest level of 60 %. The overall satisfaction of all users of the system was at a highest level of 70 %. The results could be used to reduce delays in solving problems that might occur, resulting in undamaged ready mixed concrete which could be able to be delivered on time. And it could be applied to in transportation of fuel and raw materials to production plants punctually.