BOOK OVERVIEW 9
commonly known as the strategies of exploring the “trend following” assumption for
investment. Chapter 5 introduces the principles of “follow the loser,” which is often
known as the strategies of exploring the “mean reversion” assumption for investment.
Chapter 6 introduces the principle of pattern matching for OLPS. Finally, Chapter 7
introduces the principle of meta-learning, which attempts to explore the combination
of multiple principles and strategies for OLPS.
Part III proposes four OLPS algorithms belonging to two categories, that is,
the pattern matching–based approach and follow the loser approach. The ﬁrst algo-
rithm is a pattern-matching algorithm, “CORrelation-driven Nonparametric learning”
(CORN), in Chapter 8. The other three algorithms are mean reversion algorithms.
That is, we propose the “passive–aggressive mean reversion” (PAMR) algorithm
in Chapter 9, the “conﬁdence-weighted mean reversion” (CWMR) algorithm in
Chapter 10, and the “online moving average reversion” (OLMAR) in Chapter 11.
Part IV presents our empirical studies. Chapter 12 introduces the method of empir-
ical studies, and Chapter 13 extensively evaluates the proposed algorithms on real
datasets and compares with a set of existing algorithms. Chapter 14 defends the
methodologies used in the model setting and empirical studies. Finally, Chapter 15
concludes the book with some future directions.
T&F Cat #K23731 — K23731_C001 — page 9 — 9/28/2015 — 21:04