MS: ML Models Sample Clauses

MS: ML Models. This microservice acts as a caching mechanism to avoid re-computing machine learning models in the DAL. From a data perspective (see Figure 11), it stores the models result in the appropriate entity—Classification, Cluster, or ML Model. The Classification entity stores data related to classification operations on the Feedback (e.g., feedback classification as non- functional requirement for a feature request expressed with a negative sentiment); accordingly, Cluster will store information regarding automatically identified groups of feedback (e.g., feedback concerning the same topic). Finally, ML Model stores the results of a generic machine learning model (e.g., pattern-based similarity) that operates on clusters or classes of feedback rather than on single ones. The main challenge posed by decoupling the actual machine learning algorithm from its storage mechanism (for either its input, parameters, and results) is to assess an appropriate level of abstraction. Thus, the changes necessary to synchronize this microservice with the one implementing the algorithm itself, once a new one is added to the DAL, should be minimal. Such flexibility can be achieved by using data storage that supports semi-structured data.