Identification of Structural VAR Models via Independent Component Analysis: A Performance Evaluation Study

Moneta, Alessio; Pallante, Gianluca

doi:10.1016/j.jedc.2022.104530

Cited by 8 publications

(6 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They can be used according to the proven method both for principal component analysis (PCA, [13][14][15]) and (with certain reservations) for non-negative matrix factorization (NMF, [16][17][18]). SVD can also be used to improve the results of independent component analysis (ICA, [19][20][21]). It is convenient to apply SVD because there are no restrictions on the structure of the original data matrix (square when using the LU [22] or Schur distribution [23]; square, symmetric, or positive definite when using the Cholesky distribution [24]; matrix with positive elements when applying NMF).…”

Section: Outputmentioning

confidence: 99%

“…At the same time, it was believed that these components have an abnormal distribution, and the sources of their origin are independent. To determine independent components, either minimization of mutual information based on Kullback-Leibler divergence [19] or minimization of "non-Gaussianity" [20,21] (using measures such as kurtosis coefficient and negentropy) are used. In the context of the dimensionality reduction problem, the application of ICA is trivial: to represent the input data as a mixture of components, divide them and select a certain number.…”

Section: Outputmentioning

confidence: 99%

See 1 more Smart Citation

Small Stochastic Data Compactification Concept Justified in the Entropy Basis

Kovtun,

Zaitseva,

Levashenko

et al. 2023

Entropy

View full text Add to dashboard Cite

Measurement is a typical way of gathering information about an investigated object, generalized by a finite set of characteristic parameters. The result of each iteration of the measurement is an instance of the class of the investigated object in the form of a set of values of characteristic parameters. An ordered set of instances forms a collection whose dimensionality for a real object is a factor that cannot be ignored. Managing the dimensionality of data collections, as well as classification, regression, and clustering, are fundamental problems for machine learning. Compactification is the approximation of the original data collection by an equivalent collection (with a reduced dimension of characteristic parameters) with the control of accompanying information capacity losses. Related to compactification is the data completeness verifying procedure, which is characteristic of the data reliability assessment. If there are stochastic parameters among the initial data collection characteristic parameters, the compactification procedure becomes more complicated. To take this into account, this study proposes a model of a structured collection of stochastic data defined in terms of relative entropy. The compactification of such a data model is formalized by an iterative procedure aimed at maximizing the relative entropy of sequential implementation of direct and reverse projections of data collections, taking into account the estimates of the probability distribution densities of their attributes. The procedure for approximating the relative entropy function of compactification to reduce the computational complexity of the latter is proposed. To qualitatively assess compactification this study undertakes a formal analysis that uses data collection information capacity and the absolute and relative share of information losses due to compaction as its metrics. Taking into account the semantic connection of compactification and completeness, the proposed metric is also relevant for the task of assessing data reliability. Testing the proposed compactification procedure proved both its stability and efficiency in comparison with previously used analogues, such as the principal component analysis method and the random projection method.

show abstract

Section: Outputmentioning

confidence: 99%

Section: Outputmentioning

confidence: 99%

Small Stochastic Data Compactification Concept Justified in the Entropy Basis

Kovtun,

Zaitseva,

Levashenko

et al. 2023

Entropy

View full text Add to dashboard Cite

show abstract

“…The speed at which new implementations of identification based on non-Gaussianity are developed makes comprehensive systematic comparisons difficult; current options described in Section 4 include various parametric models, moment based approaches where the researcher must choose which higher moments to use and which restrictions to impose on cross-moments, and a plethora of non-parametric estimators from the ICA literature. Moneta and Pallante (2022) is the only simulation comparison of which I am aware. There is ample scope for further and updated simulation studies comparing the leading alternatives, as well as "pre-tests" to select suitable moments containing relevant identifying variation.…”

Section: Discussionmentioning

confidence: 99%

“…For the latter, there is an important choice of which coskewness and/or co-kurtosis restrictions to impose or of which contrast function to use for ICA. Moneta and Pallante (2022) compare a variety of estimators in a simulation study. They include FastICA, the PML estimator of Gouriéroux et al (2017), and two other ICA approaches based on Givens matrices; unfortunately, they do not include recent moment-based estimators.…”

Section: Choosing the Right Functional Formmentioning

confidence: 99%

Identification based on higher moments

2024

View full text Add to dashboard Cite

“…Lu et al (2009), for instance, proposes an integration of ICA with support vector regression for forecasting of the Nikkei 225 opening index and TAIEX closing index, see also Liu and Wang (2011) and Chowdhury et al (2018). Moreover, Moneta and Pallante (2020) compare the performance of different ICA estimators within the field of structural vector autoregressive (SVAR) method on the US government spending and tax cuts data. Ceffer et al (2019) predict and model the future value of financial time series applying SVAR with ICA for pre-processing data.…”

Section: Literature Reviewmentioning

confidence: 99%

What Are the Macroeconomic Drivers of the Asset Returns of Turkish Banks?

Civan¹,

Şimşek

Çinar

2022

Technological and Economic Development of Economy

View full text Add to dashboard Cite

The aim of this paper is to investigate the effects of the macroeconomic factors to the movements of the asset returns of the banks in Turkey in terms of systemic risk from 2005 to 2018. In the study, Independent Component Analysis is applied for extracting driving factors of the asset returns of Turkish banks by decomposing the returns into its components. After examining the relationship between the independent components and the macroeconomic variables, the results conclude that one component shows strong similarities with the well-known stock market index of Turkey, namely the BIST100. Besides, the BIST100 is observed as the most important macroeconomic indicator affecting the movements of the asset returns. From systemic risk perspective, the BIST100 and the exchange rate from US dollar to Turkish lira are interpreted as two macro factors that contribute to the systemic risk of Turkish banks. When it is reviewed the regression results of the estimated independent components with the macroeconomic variables, it is found that while the BIST100 affects the asset returns of Turkish banks on its own, three macroeconomic factors (the credit default swap spreads of Turkey, the exchange rate and volatility) jointly affect the banks by creating a chain effect.

show abstract

Identification of Structural VAR Models via Independent Component Analysis: A Performance Evaluation Study

Cited by 8 publications

References 71 publications

Small Stochastic Data Compactification Concept Justified in the Entropy Basis

Small Stochastic Data Compactification Concept Justified in the Entropy Basis

Identification based on higher moments

What Are the Macroeconomic Drivers of the Asset Returns of Turkish Banks?

Contact Info

Product

Resources

About