The proposed article presents a new assessment method to inspect the relationships between financial instruments. The problem of relevant variables identification for the purpose of financial instruments valuation, as well as to assess the relationship between them, plays an important role in both, financial theories and practical investment schemes. In our approach we propose to use blind signal separation methods to decompose time series into the core components. The components possess rather physical than mathematical interpretation. Multidimensional blind source separation methods are the tool to obtain the components common to the various instruments and, together with the information about the strength of their influence, it provides broad set of characteristic describing the internal morphology of the time series.

The basic algorithm used in the research is AMUSE algorithm that accommodates second-order statistics with delays. Since the basic version of this algorithm is very sensitive to additive noise we use filtered covariance matrices in order to increase its effectiveness. We will also propose its modifications to consider the statistical relationships of higher orders. The results will be compared to other blind source separation algorithms as JADE and SOBI. Due to the fact that the blind source separation models are based on certain fundamental assumptions we will provide the broader methodological discussion regarding their application in economics.

In particular, the presented method will be applied to study the financial time series selected from different periods. The main research interest will be focused on time series morphology changes and transition states between the periods of markets collapse, and also studying the characteristics just before and after the financial crisis. We expect such analysis will deliver the insight on forthcoming market changes along with their direction and the strength.

• Legal notice:
  

  Copyright (c) Pielaszek Research, all rights reserved.
  The above materials, including auxiliary resources, are subject to Publisher's copyright and the Author(s) intellectual rights. Without limiting Author(s) rights under respective Copyright Transfer Agreement, no part of the above documents may be reproduced without the express written permission of Pielaszek Research, the Publisher. Express permission from the Author(s) is required to use the above materials for academic purposes, such as lectures or scientific presentations.
  In every case, proper references including Author(s) name(s) and URL of this webpage: https://science24.com/paper/31828 must be provided.

1. Multiway similarity approach based on divergence functions and smoothness measure.