Learning time series patterns by genetic programming

Authors:
Feng Xie;Andy Song;Vic Ciesielski
Affiliations:
RMIT University, Melbourne, Australia;RMIT University, Melbourne, Australia;RMIT University, Melbourne, Australia
Venue:
ACSC '12 Proceedings of the Thirty-fifth Australasian Computer Science Conference - Volume 122
Year:
2012

Citing 8
Cited 0

Instance-Based Learning Algorithms

Machine Learning
Genetic programming: on the programming of computers by means of natural selection

Genetic programming: on the programming of computers by means of natural selection
C4.5: programs for machine learning

C4.5: programs for machine learning
Very Simple Classification Rules Perform Well on Most Commonly Used Datasets

Machine Learning
Learning to Predict by the Methods of Temporal Differences

Machine Learning
A hybrid method for tuning neural network for time series forecasting

Proceedings of the 10th annual conference on Genetic and evolutionary computation
Liquid State Genetic Programming

ICANNGA '07 Proceedings of the 8th international conference on Adaptive and Natural Computing Algorithms, Part I
Estimating continuous distributions in Bayesian classifiers

UAI'95 Proceedings of the Eleventh conference on Uncertainty in artificial intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

Finding patterns such as increasing or decreasing trends, abrupt changes and periodically repeating sequences is a necessary task in many real world situations. We have shown how genetic programming can be used to detect increasingly complex patterns in time series data. Most classification methods require a hand-crafted feature extraction preprocessing step to accurately perform such tasks. In contrast, the evolved programs operate on the raw time series data. On the more difficult problems the evolved classifiers outperform the OneR, J48, Naive Bayes, IB1 and Adaboost classifiers by a large margin. Furthermore this method can handle noisy data. Our results suggest that the genetic programming approach could be used for detecting a wide range of patterns in time series data without extra processing or feature extraction.