Evaluation of consumer comments made on e-commerce websites by artificial intelligence

E-commerce is a digital commercial model that encompasses online buying and selling transactions. With the widespread adoption and increased accessibility of the internet, e-commerce plays an indispensable role in today's business world, offering advantages such as reaching a broader customer base, reducing costs, and increasing transaction speed. Consumer reviews on e-commerce websites are considered significant marketing tools for businesses. It is acknowledged that positive reviews can boost sales, while negative reviews may lead to potential customer loss. Therefore, e-commerce platforms need to consider, effectively manage, and focus on feedback to enhance customer satisfaction. This study aims to obtain meaningful insights from online customer product reviews and explore customer sensitivity through a proposed computational model leveraging artificial intelligence. To achieve this goal, an open- access Turkish dataset containing a total of 2000 consumer reviews was created. A deep learning model based on Long Short-Term Memory (LSTM) and a novel text encoding method were proposed to automate the analysis of consumer reviews. The supervised learning approach resulted in a model achieving a 91.50% accuracy rate. This high accuracy suggests that an artificial intelligence model with such performance can produce meaningful results on product pages containing hundreds or thousands of customer reviews. The study highlights the potential use of AI-supported algorithms for businesses to enhance customer satisfaction, develop marketing strategies, manage online marketing processes, and implement various business models.