How does this work technically?
Unsupervised
Recommendation engine algorithms require little or no training data, so they are unsupervised algorithms. The value that must be “recommended” is taken from the nearest data point. This means that the distance function used to calculate this point is a crucial part. As with other AI and Deep Learning algorithms, various functions and configurations are possible, whether or not combined with additional clustering or classification algorithms. The evaluation of these combinations is carried out independently by Trendskout Auto ML .
Recommendation vs Classification
Recommendation engines are used when there are many different types of content, products or services that are eligible to present. If this number is limited, a Classification analysis is more appropriate, for which a training phase is required.
