Implicit Fusion¶
Implicit Fusion refers to joining data sources using information implicit within each. Under one framework, Statistical-Matching (Hot Deck Imputation), overlapping features are used to determined how simlilar records are, and assignments/matches are made using various algorithms. Under another method, a model is first built to predict the target(s). Predictions are made for both data sources, and they are matched based on these. This method is sometimes referred to as Predictive Mean Matching. These are the core implicit algorithms most often used in the data fusion literature and applications, and are the only ones currently offered.