Support Spinor Machine

Kabin Kanjamapornkul,Richard Pinčák,Sanphet Chunithpaisan,Erik Bartoš

DOI: https://doi.org/10.1016/j.dsp.2017.07.023

2017-09-11

Abstract:We generalize a support vector machine to a support spinor machine by using the mathematical structure of wedge product over vector machine in order to extend field from vector field to spinor field. The separated hyperplane is extended to Kolmogorov space in time series data which allow us to extend a structure of support vector machine to a support tensor machine and a support tensor machine moduli space. Our performance test on support spinor machine is done over one class classification of end point in physiology state of time series data after empirical mode analysis and compared with support vector machine test. We implement algorithm of support spinor machine by using Holo-Hilbert amplitude modulation for fully nonlinear and nonstationary time series data analysis.

Machine Learning,Signal Processing,Statistical Finance

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the limitations of the existing Support Vector Machine (SVM) in dealing with nonlinear and non - stationary time - series data, especially financial time - series data. Specifically, SVM is based on the Euclidean plane and the Hausdorff separation axiom (T2), which has algebraic structure flaws when dealing with certain types of input data and cannot effectively classify data with topological entanglement characteristics.

To solve these problems, the author introduced a new machine - learning model - Support Spinor Machine (SSM). SSM improves SVM in the following ways:

1. **Expand feature space**: Generalize the vector field to the spinor field and extend the separating hyperplane to time - series data in Kolmogorov space, so that it can handle higher - dimensional mathematical objects.
2. **Use T0 separation axiom**: Different from the T2 separation axiom used by SVM, SSM uses the T0 separation axiom to deal with time - series data on non - Euclidean planes, so it can better handle data with topological defects.
3. **Introduce modular space model**: Use the modular space model to classify time - series data, which helps to capture the complex structures and dynamic changes in the data.
4. **Apply holographic Hilbert modulation algorithm**: Use holographic Hilbert amplitude modulation (Holo - Hilbert Amplitude Modulation) to analyze fully nonlinear and non - stationary time - series data, improving the flexibility and adaptability of the model.

Through these improvements, SSM can more effectively deal with nonlinear and non - stationary time - series data, especially showing better performance in financial time - series prediction. The author also verified the superiority of SSM in one - day direction prediction through empirical analysis and compared it with the traditional SVM.