Evaluating the quality of manufactured products by providing an approach based on the ANFIS neural-fuzzy network (case study: Khazar Plastic Manufacturing and Industrial Company)

Document Type : Original Article


1 PhD student, Department of Management, Faculty of Management and Economics, University of Guilan, Guilan, Iran.

2 Associate Professor, Department of Civil Engineering, Technical and Engineering Faculty, University of Guilan, Guilan, Iran.

3 Associate Professor, Department of Management, Faculty of Management and Economics, University of Guilan, Guilan, Iran.



Currently, the need to expand accurate and fast quality assurance to provide high quality and safe manufactured products; Yes, to achieve this goal in the production industry, there are problems such as low production efficiency, low accuracy, and lack of innovation in products. A challenge for traditional methods in quality management is the management of high-dimensional and non-linear production data. In this research, an adaptive neuro-fuzzy model, which has the advantages of neural network and fuzzy inference, to evaluate and extract the quality level of manufactured products. In order to train the proposed model, the collection of data related to the quality process of one part of the production line, Khazar Plastic Industrial Company, and 550 data related to the quality process with an emphasis on influential variables such as "appearance standard, external diameter of the large gear wheel, thickness of the gear wheel, shaft length Metal, height of metal shaft, outer diameter of metal shaft" is considered as input variable and final quality as output variable. Finally, the accuracy of the results and the effectiveness of the proposed adaptive neuro-fuzzy model have been evaluated using the statistical indices of correlation coefficient and root-mean-square error. The results show the data used to evaluate the quality of the production part in the adaptive-proposal neuro-fuzzy method With a correlation coefficient of 0.95 and a mean square error of 0.42869, it provided a good match between the quality of the model output and the actual values


Main Subjects