Enhancing Product Categorization in E-commerce using NLP and Machine Learning

E-commerce, or electronic commerce, has revolutionized the way businesses operate, and consumers engage in transactions, also having a profound impact on global economies. From the early days of online retail to the current era of seamless digital transactions, the evolution of e-commerce has been marked by innovations in payment systems, security protocols, and user experience. In this work, the key problem addressed is product categorization based on the text description of a product in a particular category. Natural language processing techniques are used to eliminate unwanted and irrelevant information from the text description of the products. The normalized text description is then vectorized using the Term Frequency-Inverse Document Frequency method. This vectorized data is input into the machine learning models with fine-tuned hyperparameters to classify into different product categories. Support vector machines resulted in the highest accuracy in classifying the product categories, which is 95.19%. By accurately categorizing products, it becomes easier for users to find the relevant products, improving the overall user experience. This can lead to increased customer satisfaction and potentially higher sales.