Particle Swarm Optimization of Bollinger Bands

Abstract

Abstract. The use of technical indicators to derive stock trading signals is a foundation of financial technical analysis. Many of these indicators have several parameters which creates a difficult optimization problem given the highly non-linear and non-stationary nature of a financial timeseries. This study investigates a popular financial indicator, Bollinger Bands, and the fine tuning of its parameters via particle swarm optimization under 4 different fitness functions: profitability, Sharpe ratio, Sortino ratio and accuracy. The experiment results show that the parameters optimized through PSO using the profitability fitness function produced superior out-of-sample trading results which includes transaction costs when compared to the default parameters.

A learning adaptive Bollinger band system

“…The quality of these solutions will dictate the effectiveness of the overall strategy. Based on the implication of cyclical profitability as described by Lo in the adaptive market hypothesis [10] and the results from Butler et al [2] which demonstrate this property, this study generates an initial population of ABBs that in theory should be profitable, on average, in 20% of the out-of-sample data 1 . Therefore this approach is dependent on the ability of a meta agent to identify ABBs that will generate the best estimate of future market behaviour.…”

“…The contributions of this paper are: (1) to provide a framework that extends the research of optimizing technical analysis with AI techniques to a combined signal approach, (2) extend an existing Particle Swarm Optimization algorithm to a multiobjective optimization, (3) to augment the LCS framework for online learning when the population of solutions are not classifiers, and (4) to extend the idea of cyclical profitability from the adaptive market hypothesis (AMH) into a whitebox forecasting tool. The following sections will describe the various components of the Learning Adaptive Bollinger Band System (LABBS) and review its forecasting performance against some relevant benchmarks.…”

Testing implications of the Adaptive Market Hypothesis via computational intelligence

“…This would entail determining the length of windows for calculating the moving averages via profitability based fitness functions. Another recent study combined Bollinger Bands with Particle Swarm Optimization (PSO) [1] to tune the parameters to current market conditions. The experiments implied that the effectiveness of the indicator could be enhanced beyond that of just using the default parameters.…”

“…For this study the optimal trader for each market segment will be determined using Adaptive Bollinger Bands (ABB) [1], which are based on a technical indicator created by John Bollinger in the 1980's.…”

“…A result from [1] concluded that BBs are ineffective at generating profits when the market is trending. This shortcoming of the BBs was mainly due to the exiting of profitable trades prematurely.…”