Calculation of the productivity of construction gypsum plaster worker using support vector machine algorithm

Abstract

Calculating the productivity of workers in traditional construction projects can be a daunting and time-consuming task. However, thanks to the advancements in technology and scientific research, measuring worker productivity can be automated. Therefore, this study aims to determine the productivity of gypsum plaster workers by collecting motion data with a sensor on their arms and using the support vector machine algorithm for analysis. Based on the estimation results, motion productivity is determined by calculating the ratio of the worker's working time to the total time. On the other hand, worker productivity is calculated by determining the ratio of the amount of work completed to the product of the number of workers and the total working hours. Finally, theoretical productivity is calculated by dividing motion productivity by the calculated worker productivity. According to the results of the analysis, whether the worker is working or not has been estimated correctly by 95.8%. On the other hand, the mean daily theoretical productivity has been determined to be 10.20 m 2 /man-hour. In this way, workers' activities can be automatically detected with a certain accuracy and their productivity can be calculated. This helps in the effective management of the construction site.