Lead-lag detection and network clustering for multivariate time series with an application to the US equity market

Bennett, S. M.; Cucuringu, Mihai; Reinert, Gesine

doi:10.48550/arxiv.2201.08283

Cited by 3 publications

(2 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Explaining preference models is crucial when they are applied in areas such as recommendation systems [9], finance [10], and sports science [11] for the practitioner to trust, debug and understand the value of their findings [12]. However, despite its importance, no prior work has studied this problem to the best of our knowledge.…”

Section: Introductionmentioning

confidence: 99%

Explaining Preferences with Shapley Values

Hu¹,

Chau²,

Huertas³

et al. 2022

Preprint

View full text Add to dashboard Cite

While preference modelling is becoming one of the pillars of machine learning, the problem of preference explanation remains challenging and underexplored. In this paper, we propose PREF-SHAP, a Shapley value-based model explanation framework for pairwise comparison data. We derive the appropriate value functions for preference models and further extend the framework to model and explain context specific information, such as the surface type in a tennis game. To demonstrate the utility of PREF-SHAP, we apply our method to a variety of synthetic and real-world datasets and show that richer and more insightful explanations can be obtained over the baseline. * equal contribution, order decided by coinflip Preprint. Under review.

show abstract

Section: Introductionmentioning

confidence: 99%

Explaining Preferences with Shapley Values

Hu¹,

Chau²,

Huertas³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Their method derives from spectral clustering and different in nature: In their case the network structure of the data is unobservable and the clustering is applied to the partial correlation structure, whereas in our case the clustering is directly applied to the the network structure of firms, assumed heterogeneous. Other works on equity time series construct a directed network using lead-lag relationships Bennett et al (2022) and reside on novel spectral network clustering methods Cucuringu et al (2020).…”

mentioning

confidence: 99%

Clustering heterogeneous financial networks

Amini,

Chen,

Minca

et al. 2023

Mathematical Finance

View full text Add to dashboard Cite

We develop a convex‐optimization clustering algorithm for heterogeneous financial networks, in the presence of arbitrary or even adversarial outliers. In the stochastic block model with heterogeneity parameters, we penalize nodes whose degree exhibit unusual behavior beyond inlier heterogeneity. We prove that under mild conditions, this method achieves exact recovery of the underlying clusters. In absence of any assumption on outliers, they are shown not to hinder the clustering of the inliers. We test the performance of the algorithm on semi‐synthetic heterogenous networks reconstructed to match aggregate data on the Korean financial sector. Our method allows for recovery of sub‐sectors with significantly lower error rates compared to existing algorithms. For overlapping portfolio networks, we uncover a clustering structure supporting diversification effects in investment management.

show abstract

Graph similarity learning for change-point detection in dynamic networks

Sulem,

Kenlay,

Cucuringu

et al. 2023

Mach Learn

View full text Add to dashboard Cite

Dynamic networks are ubiquitous for modelling sequential graph-structured data, e.g., brain connectivity, population migrations, and social networks. In this work, we consider the discrete-time framework of dynamic networks and aim at detecting change-points, i.e., abrupt changes in the structure or attributes of the graph snapshots. This task is often termed network change-point detection and has numerous applications, such as market phase discovery, fraud detection, and activity monitoring. In this work, we propose a data-driven method that can adapt to the specific network domain, and be used to detect distribution changes with no delay and in an online setting. Our algorithm is based on a siamese graph neural network, designed to learn a graph similarity function on the graph snapshots from the temporal network sequence. Without any prior knowledge on the network generative distribution and the type of change-points, our learnt similarity function allows to more effectively compare the current graph and its recent history, compared to standard graph distances or kernels. Moreover, our method can be applied to a large variety of network data, e.g., networks with edge weights or node attributes. We test our method on synthetic and real-world dynamic network data, and demonstrate that it is able to perform online network change-point detection in diverse settings. Besides, we show that it requires a shorter data history to detect changes than most existing state-of-the-art baselines.

show abstract

Lead-lag detection and network clustering for multivariate time series with an application to the US equity market

Cited by 3 publications

References 52 publications

Explaining Preferences with Shapley Values

Explaining Preferences with Shapley Values

Clustering heterogeneous financial networks

Graph similarity learning for change-point detection in dynamic networks

Contact Info

Product

Resources

About