Is content-based recommendation algorithm unsupervised?

Content-based recommendation algorithm is used as an example of unsupervised learning. However, if you look at the details of the algorithm, it trains deep learning. So, I checked if the algorithm is supervised or unsupervised.

In a nutshell, it uses both. In the context of recommendation systems, training a deep learning model, such as an artificial neural network (ANN), is typically a supervised learning process. It involves using labeled data, such as user-item interactions or explicit feedback, to learn the relationship between input features and target variables (e.g., ratings or preferences).

However, once the deep learning model is trained, you can use the output of the model to represent items or users in a latent feature space. This is where the unsupervised learning aspect comes into play. By calculating the similarity between items (or users) based on their latent features, you can generate recommendations without requiring explicit supervision or labeled data.

The process of calculating similarities between items (or users) based on the output of a deep learning model can be seen as an unsupervised learning task. This is because it involves discovering patterns and relationships between items (or users) without using labeled data or explicit feedback.

In summary, training a deep learning model for recommendation systems is typically a supervised learning process, while using the output of the deep learning model to calculate similarities between items (or users) and generate recommendations is an unsupervised learning task. This combination of supervised and unsupervised learning techniques can help create more effective and personalized recommendation systems.