| Caner Kıvanç Hekimoğlu, Mehmet Süleyman Ünlütürk, Aybüke Keçeci, Muhterem Kamil Bilir Application of a Fashion Consultant Software Utilizing Neural Networks in Ready-to-wear Cloth Retailing |
|---|
| Abstract. A software application utilizing a multilayer feedforward neural network developed in this study can be effectively used as a fashion consultant within ready-to-wear cloth retailers. The designer uses a mobile phone application to interact with the neural network system and this mobile phone application calculates the unique features associated with the candidate garment design and sends them to the multilayer feedforward neural network that runs inside the server. The designer receives an acceptance rate as a response displayed on the mobile phone application within seconds and decides whether to send the candidate garment design drawing to the production department based on this prediction. The calculated rate was generated by the multilayer feedforward neural network which was trained on the client’s history of purchases. 87% accuracy was achieved with this neural network during the testing. For each garment design drawing, there are a lot of man-hours spent in the ready-to-wear cloth production department to create the actual garment. If the neural network can predict successfully that the actual design will be rejected when it is received by the client, then the man-hours that will be spent in the production department for this design will be saved. |
| Keywords: Multilayer Feedforward Neural Network, Mobile Phone Application, Digital Image Processing, Fashion Consultant |
 Download PDF |
| DOI: https://doi.org/10.54381/itta2024.17 |