Analysis of comments on e-commerce sites with bert and albert language models

Online shopping (e-commerce) platforms, which have become popular today, are often preferred by many people because they offer a fast, easy, secure and contactless shopping experience. Positive or negative comments made on online platforms about the product and service that customers apply for, in order to enable them to make the right decision before making the purchase during online shopping, play an important role in their shopping preferences. Customer reviews of products on e-commerce platforms help customers get a realistic idea about the product/service and help businesses improve their performance by evaluating these comments. However, reading and analyzing these comments one by one by customers is a time-consuming and difficult process. This process also brings the risk of human error in the analysis of product reviews due to effects such as distraction. Various studies are carried out on the development of language analysis models based on machine learning algorithms in order to enable customers to make fast and accurate decisions about the product and service they are interested in. In this study, it is planned to conduct a sentiment analysis about the product/service by using customer comments on products/services on e-commerce sites using BERT and ALBERT language models. It is predicted that the results of the analysis obtained in the study will increase customer satisfaction on e-commerce sites and contribute to issues such as marketing strategies and product development.