The machine learning approach is determined by the nature of the data available and the desired output.
Data Available: Customer reviews (input) that are manually tagged with a sentiment category (output/label).
Desired Output: A model that can classify new, untagged reviews into one of the predefined categories (positive, negative, or neutral).
This scenario perfectly aligns with the definition of Supervised Learning (D). Supervised learning is the machine learning paradigm where the model is trained on a labeled dataset—a dataset where the input data is explicitly paired with the correct output label. The model learns a function that maps the input (the review text) to the output (the sentiment tag) and is then used to predict the label for unseen data.
Unsupervised Learning (B) is used for unlabeled data to find hidden patterns or groupings (clustering), which is not the goal here.
Reinforcement Learning (C) is used for training an agent through trial and error using a system of rewards and penalties.
Deep Learning (A) is a type of model (using deep neural networks) that can be used for supervised learning, but the learning approach required here is definitively supervised.
(Reference: Google's training materials on Machine Learning Approaches define Supervised Learning as training a model using labeled data to make predictions or classifications for new, unseen inputs. Sentiment analysis is a canonical example of a supervised learning classification task.)
