Gated Spiking Neural P Systems for Time Series Forecasting

Area of Science:

• Computational Neuroscience
• Artificial Intelligence
• Time Series Analysis

Background:

• Spiking neural P (SNP) systems are neural-like computing models inspired by spiking neurons.
• Existing SNP systems have limitations in controlling neuron state dynamics.

Purpose of the Study:

• To introduce a new variant of SNP systems called gated spiking neural P (GSNP) systems.
• To develop a GSNP-based model for effective time series forecasting.

Main Methods:

• Introduced gated neurons with reset and consumption gates to control neuron state updates.
• Developed the GSNP model for time series prediction.
• Evaluated the GSNP model on benchmark univariate and multivariate time series datasets.

Main Results:

• The GSNP model demonstrated effective control over neuron state dynamics.
• Comparative analysis showed the GSNP model outperforms several state-of-the-art prediction models.
• The proposed GSNP system proved effective for time series forecasting.

Conclusions:

• GSNP systems offer a promising advancement in neural-like computing models.
• The GSNP model provides a robust and effective solution for time series forecasting challenges.