Identification of Customer Facial Expressions to Improve Service Experience Using Convolutional Neural Networks (CNN)

Abstrak

Service experience is an important factor that determines the success and sustainability of a business. Understanding customer facial expressions can help business owners identify emotional responses that arise during the service process. This study develops a model for identifying customer facial expressions using a Convolutional Neural Network (CNN) with the MobileNetV2 architecture to enhance the customer service experience at Toko Saudara. This study aims to develop a system model that can identify customer facial expressions and measure the accuracy of the model in classifying facial expressions based on the dataset used. The model was trained using a transfer learning approach with training parameters of 50 epochs, a batch size of 32, and a learning rate of 0.0001 using the Adam optimizer. The dataset consisted of 2000 augmented images from customer CCTV recordings with two main classes, namely satisfied and dissatisfied. The dataset was split into 80% training, 10% validation, and 10% testing portions. The model achieved a training accuracy of 95.94% with a loss value of 0.1377, and testing performance reached 98% accuracy, with precision, recall, and F1-score all at 98%. The resulting model is able to accurately identify customer facial expressions so that it can be used by businesses to understand customer emotional responses and improve service quality.

Identifier

• OpenAlex: https://openalex.org/W4416455270
• DOI: https://doi.org/10.24167/sisforma.v12i2.14314