Summary

We covered a lot of ground in this chapter. Focusing on the sentiment analysis problem using real-world reviews from the Amazon online store, we became better acquainted with different algorithms and methods for supervised learning. Simultaneously, we broadened our coverage on how algorithms learn from data and how to incorporate optimization techniques for this task.

We worked on more advanced plots, starting with the EDA phase, and provided both cumulative and individual statistics for the reviewers. Additionally, we found an indirect way to assign a sentiment label to the data samples utilizing the reviewers’ ratings.

The discussion around logistic regression facilitated the introduction of avoiding overfitting using regularization. Then, we detailed how artificial neurons are networked together to form complex networks. Finally, both algorithms were used to classify the samples in the dataset and provided good performance. Up next, we have another problem to...