Non-Parametric Time Series Classification

We present an improved state-based prediction algorithm for time series. Given time series produced by a process composed of different underlying states, the algorithm predicts future time series values based on past time series values for each state. Unlike many algorithms, this algorithm predicts a multi-modal distribution over future values. This prediction forms the basis for labelling part of a time series with the underlying state that created it given some labelled example signals. The algorithm is robust to a wide variety of possible types of changes in signals including changes in mean, amplitude, amount of noise, and period. We show results demonstrating that the algorithm successfully segments signals from several robotic sensors generated while performing a variety of simple tasks.

Authors

• Scott Lenser
• Manuela Veloso

Keywords

• Robot sensing systems
• Prediction algorithms
• Intelligent sensors
• Intelligent robots
• Signal generators
• Signal processing
• Hidden Markov models
• Predictive models
• Labeling
• Noise robustness
• Time Series
• Time Series Classification
• Types Of Changes
• Future Values
• Wide Variety Of Types
• Time Series Of Values
• Robot Sensors
• Probability Density
• Series Data
• Output Value
• Bright Light
• Accelerometer Data
• Time Series Prediction
• Previous Series
• Finite Bandwidth
• probabilistic models
• sensors
• Markov models