9.5.1 Fellegi and Sunter Classifier

The classical Fellegi and Sunter classifier [13] simply sums all the log2-weights in a weight vector (as calculated by a RecordComparator as discussed in Section 9.3), and then uses two thresholds to classify a record pair into one of the three classes links, non-links or possible links. The results of a classification are stored in a data structure, which can then be used to produce various output forms as presented in Chapter 11.

When a Fellegi and Sunter classifier is initialised, the following arguments need to be given.

name
A name for a the classifier. This should be a short string.
dataset_a
A reference to a data set. This must be the same data set as the first data set defined within a record comparator.
dataset_b
A reference to a data set, which must be the same as the second data set defined within a record comparator. This data set will be the same as dataset_a in a deduplication process, but most likely a different data set in a linkage process (until different parts of the same data set are to be linked).
lower_threshold
A number, which is the lower threshold for the classifier.
upper_threshold
A number, which is the upper threshold for the classifier. It must be equal to or larger than the lower threshold.

The following example shows how a Fellegi and Sunter classifier can be initialised.

# ====================================================================

f_s_classifier = FellegiSunterClassifier(name = 'My F & S classifier',
                                    dataset_a = mydata_1,
                                    dataset_b = mydata_2,
                              lower_threshold = 10.0,
                              upper_threshold = 50.0)