In this chapter, we learned various NLP techniques, namely BoW, Word2vec, GloVe, and fastText. We built projects involving these techniques to perform sentiment analysis on an Amazon reviews dataset. The projects that were built involved two approaches, making use of pretrained word embeddings and building the word embeddings from our own dataset. We tried both these approaches to represent text in a format that can be consumed by ML algorithms that resulted in models with the ability to perform sentiment analysis.

In the next chapter, we will learn about customer segmentation by making use of a wholesale dataset. We will look at customer segmentation as an unsupervised problem and build projects with various techniques that can identify inherent groups within the e-commerce company's customer base. Come, let's explore the world of building an e-commerce customer segmentation engine with ML!