Developing a hybrid intelligent model for constructing a size recommendation expert system in textile industries

Abstract

Purpose – In shopping, for selecting the appropriate garments, people have to try on multiple garments. This problem is due to lack of a sizing system based on updated anthropometric data and the classification system that introduces the appropriate size from the sizing chart to each person. To solve this problem, as a first study in the literature, a hybrid intelligent classification model as a size recommendation expert system is proposed. The paper aims to discuss these issues. Design/methodology/approach – Three stages for developing a hybrid intelligent classification system based on data clustering and probabilistic neural network (PNN) are proposed. In the first stage, the clustering algorithm is used for specifying the sizing chart. In the second stage, the resulting sizing chart is used as a reference for developing a new intelligent classification system by using a PNN. At the last stage, the accuracy of the proposed model is evaluated by using the Iranian male's body type data set. Findings – Experimental results show that the proposed model has a good accuracy and can be used as a size recommendation expert system to specify the right size for the customers. By using the proposed model and designing an interface for it, a decision support system was developed as a size recommendation expert system that was used by an apparel sales store. The results were time saving and more satisfying for the customers by selecting the appropriate apparel size for them. Originality/value – In this paper, as a first study in literature, a hybrid intelligent model for developing a size recommendation expert system based on data clustering and a PNN to enable the salesperson to help the consumer in choosing the right size is proposed. In the first stage, the clustering algorithm is used for specifying the sizing chart. In the second stage, the resulting sizing chart is used as a reference to develop a new intelligent classification system by using a PNN. In the last stage, the accuracy of the proposed model is evaluated by using testing data. The proposed model achieved an 87.2 percent accuracy rate that is very promising.