Quality improvement through the preventive detection of potentially defective products in the automotive industry by means of advanced artificial intelligence techniques

Abstract

The paper addresses the problem of quality assessment of high-precision automotive components by means of artificial intelligence techniques that aim at the detection of potentially defective products before they are sold to customers. This control is motivated by industrial requirements as it could avoid a number of negative consequences for the company. The problem involves the classification of strongly unbalanced datasets and requires a suitable preprocessing of the data before they are used for models training. Standard classifiers and ensemble methods were used for the detection of defective products. The satisfactory results achieved by the selected approach lead to significant improvement from the industrial point of view.