Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Marmara Üniversitesi, Fen Bilimleri Enstitüsü, Türkiye
Tezin Onay Tarihi: 2021
Tezin Dili: İngilizce
Öğrenci: AHMET KUTALMIŞ COŞKUN
Danışman: Mustafa Borahan Tümer
Özet:
Dynamic systems are highly complex and hard to deal with due to their subject-
and time-varying nature. The fact that most of the real world systems/events are
of dynamic character makes modeling and analysis of such systems inevitable and
charmingly useful. Working with such systems require the learning algorithm to be
able to adapt to new conditions that might occur at unknown times.
One promising estimation method that is capable of unlearning past information
to deal with non-stationarity is Stochastic Learning Weak Estimator (SLWE) by
Oommen and Rueda (2006). This method maintains an always up-to-date estimate of
the target parameter by employing a multiplicative update rule that adjusts previous
estimate based on the current observation. However, due to using a constant learning
rate, it faces a trade-off between plasticity and stability, which is also referred to as
exploration versus exploitation dilemma in some contexts.
In this study, considering the problem of estimating parameters that characterize a
statistical distribution that involves abrupt changes, SLWE is modeled as a random
walk and asymptotic behavior of estimates is theoretically analyzed to obtain a
statistical model. Utilizing this model, changes in stationarity is detected to switch
between exploratory and exploitative learning modes. Experimental evaluations on
both synthetic and real world data show that the proposed method tracks the target
parameter with 15% to 50% less error.