Analysis of e-commerce product reviews with machine learning

Today, with the development of technology, the e-commerce sector has grown very rapidly. With the development of the e-commerce sector, the comments made on the products have increased tremendously. As the number of reviews of the products increased, it became very difficult to examine and analyse them one by one, and the sellers became unable to evaluate all the comments made on their products. In the face of these problems, the need to create an artificial intelligence supported automation has emerged. In this study, a method has been developed for the classification of product reviews, which is seen as the first step of automation. First of all, 15,170 product reviews were collected from e-commerce platforms operating in our country, and these comments were labelled as positive, negative and neutral twice at different times, and a two-class data set was prepared by removing neutral comments. With this data set, a model was developed using the Long-Short Term Memory algorithm, and with this model, the data set containing 203,274 comments collected from e-commerce platforms was automatically labelled. In this way, experiments were carried out with machine learning and deep learning algorithms using two manually and automatically labelled data sets. As a result of the study, very high results were obtained in the experiments in which the automatically labelled data set with the developed model was used. By developing a web-based application in which the classification model we have created is used, we have provided solutions to the problems of automatic classification of product reviews and automatic labelled dataset creation in academic studies.