"Variants of Recursive Consequent Parameters Learning in Evolving Neuro-Fuzzy Systems"
: Proceedings of the IEEE EAIS 2020 Conference, Serie Proceedings of the IEEE EAIS 2020 Conference, IEEE Press, 2020
Variants of Recursive Consequent Parameters Learning in Evolving Neuro-Fuzzy Systems
Sprache des Titels:
Proceedings of the IEEE EAIS 2020 Conference
During the last 15 to 20 years, evolving (neuro-
)fuzzy systems (E(N)FS) have enjoyed more and more attraction
in the context of (fast and real-time) data stream mining and
modeling processes. This is because they can be updated on the
fly in a single-pass sample-wise manner and are able to perform autonomous changes on structural level in order to account for drifts or dynamic extensions of process ranges. A wide variety of approaches have been proposed to handle these issues by updating the rule structure and antecedents. The current denominator in the update of the consequent parameters is the usage of the recursive (fuzzily weighted) least squares estimator (R(FW)LS), as being applied in almost all E(N)FS approaches so far. In this paper, we propose and examine alternative variants for consequent parameter updates, namely multi-innovation RFWLS, recursive corr-entropy and especially recursive weighted total least squares.
Multi-innovation RLS guarantees more stability in the update,
whenever structural changes (i.e. changes in the antecedents) in the E(N)FS are performed, as the rule membership degrees on
(a portion of) past samples are actualized before and properly
integrated in each update step. Recursive corr-entropy addresses the problematic of outliers by down-weighing the influence of (atypically) higher errors in the parameter updates. Recursive weighted total least squares takes into account also a possible noise level in the input variables (and not solely in the target variable as in RFWLS). The approaches are compared with standard RFWLS i.) on three data stream regression problems from practical applications, affected by (more or less significant) noise levels and one embedding a known drift, and ii.) on a real-world time-series based forecasting problem, also affected by noise and where the EFS is evolved and updated in an LVinduced
score space. The results based on accumulated prediction
error trends over time indicate that RFWLS is not necessarily the best choice for consequent parameter updating. So, the proposed variants could be worth of being further considered as promising and serious alternatives.