https://wiki.umiacs.umd.edu/clip/ngfci/index.php?title=Special:NewPages&feed=atom&hideredirs=1&limit=50&offset=&namespace=0&username=&tagfilter=&size-mode=max&size=0ngfci - New pages [en]2024-03-29T08:52:21ZFrom ngfciMediaWiki 1.39.6https://wiki.umiacs.umd.edu/clip/ngfci/index.php/KarshaVizAnKarshaVizAn2016-05-19T00:46:39Z<p>Louiqa: </p>
<hr />
<div>[[Media:[[Media:Example.ogg]]]] The Karsha Explorer tool can visualize monthly price variations for equities in the S&P 500 Index. <br />
The user can filter the monthly data to highlight equities that experience significant variations. <br />
The tool will allow the user to compare signals from individual equities that have experienced <br />
significant variations with the contemporaneous moves in the S&P 500 index for that month. <br />
These combinations of extreme moves in individual equities and the market, together with turnover<br />
and other features are the basis to create pattern templates.<br />
<br />
Locally Important Maxima and Minima LIMx and LIMn [http://karsha.umiacs.umd.edu/karshaCEP/Testing.html] [http://karsha.umiacs.umd.edu/karshaCEP/PRC_analysis.html]<br />
<br />
Monthly Drawdowns [http://karsha.umiacs.umd.edu/Drawdowns/]<br />
Karsha Explorer short paper [[file:KarshaExplorer.pdf]]</div>Louiqahttps://wiki.umiacs.umd.edu/clip/ngfci/index.php/CorpDebtRepCorpDebtRep2016-05-03T22:36:47Z<p>Louiqa: </p>
<hr />
<div>Corporations play a significant role in our economic and financial eco-systems. Corporate debt is often a key driver of important business interactions and transactional activity. While it represents a significant fraction of US debt, there is limited public knowledge about corporate debt instruments, and the role played by financial institutions along the corporate debt supply chain. While financial researchers often have a clear theoretical understanding of how debt structure, cash flows, covenants, etc. can impact liquidity measures or can lead to contagion that can potentially increase systemic risk, there are currently few public resources available to enable rigorous data-driven systemic studies of corporate debt.<br />
<br />
This research will address the “data science for finance” challenges of creating a public CorpDebt repository <br />
that will provide precise yet detailed insights into the corporate debt eco-system, capturing relationships <br />
among financial institutions, and specific details about a corporation and its debt structure. <br />
<br />
This is an important first step in creating a representation of the complete debt market, which includes mortgage backed, other asset backed, municipal, federal and sovereign debt products. The intellectual merit of this research is to harness computational methods and financial big data to create rich and relevant resources and tools that can be exploited by financial researchers. <br />
<br />
We will utilize cloud based text analytics and machine learning, ontological knowledge, crowd-sourced wisdom, <br />
and financial big data from prospecti, to construct an open, rich and complex repository, CorpDebt, centered <br />
around US corporate debt. A description of the Corporate Debt Repository project is available here [https://drive.google.com/open?id=0BzTeYQSh4QTsa2hPQjdCcWJ6VXc].<br />
<br />
The CorpDebt repository will build upon the following ontological concepts defined in the Financial Industry Business Ontology (EDM Council 2016):<br />
<br />
* Relationships: e.g., a focal corporation, a parent or subsidiary, affiliate, etc.<br />
* Guarantees, covenants, corporate events, etc.<br />
* Amount and timing of new issuances, debt structure, seniority, cash flow, etc.<br />
<br />
Current resources, e.g., Bloomberg, will typically identify the existence of a type of covenant and provide a link to the text on that page but the resource will not process the language of the covenant, identify events, etc.<br />
<br />
Impact on financial stability and monitoring: The CorpDebt repository will enable academic researchers and regulators to undertake systemic data-driven research, at a deep(er) level of granularity about debt products, and incorporating semantic knowledge about corporate relationships; such research would not be an option using traditional resources such as COMPUSTAT. Research hypotheses could be validated by combining data from the CorpDebt repository with additional time-series econometric and financial datasets. For example, researchers studying corporate debt liquidity or contagion can use the CorpDebt repository to create panel datasets across various industry sectors, market capitalization levels, etc. They can also include network-level metrics reflecting connected communities of financial institutions, concentration, outliers, etc. CorpDebt may also be able to discover, and provide a priori alerts about evolving and emerging (latent) relationships among systemically important financial institutions. The CorpDebt repository will thus provide regulators with high quality knowledge; it will improve transparency and it will provide an additional valuable resource to monitor systemic risk.</div>Louiqahttps://wiki.umiacs.umd.edu/clip/ngfci/index.php/ResMBSResMBS2016-04-13T02:30:18Z<p>Louiqa: </p>
<hr />
<div><br />
resMBS is a graph / dataset that has been extracted from the contents of financial prospecti for US <br />
residential mortgage backed securities filed with the SEC. These securities started becoming very<br />
popular in 2002. The issued securities reached a peak in 2006 and then started to decline in 2007 <br />
and came to an abrupt end in 2008. <br />
<br />
We extracted the "financial supply chain" comprising "financial institutions" (FI) and the role (Role) <br />
that they play on a financial contract (FC).<br />
<br />
The following paper provides an overview of how the dataset was created and some preliminary clustering <br />
analysis on the graph.<br />
resMBS: Constructing a Financial Supply Chain from Prospecti<br />
Doug Burdick, IBM<br />
Soham De and Louiqa Raschid and Mingchao Shao and Zheng Xu and Elena Zotkina, University of Maryland<br />
[https://drive.google.com/open?id=0BzTeYQSh4QTsWmJXSTM0TEFLa1E]<br />
<br />
The networks described in the paper can be viewed here.<br />
FI clusters [http://dsfin.umiacs.umd.edu/FI-network/4roles/]<br />
FI clusters based on FC-FC similarity [http://dsfin.umiacs.umd.edu/FC-network/4roles/]<br />
FI-FC bipartite graph [http://pattaran.umiacs.umd.edu/clusters/resmbs-4roles]<br />
<br />
We used a topic modeling approach to develop a model FI-Comm where a topic is defined over a vocabulary<br />
of FIs and a model Role-FI-Comm where a topic is defined over a vocabulary of Role-FI pairs.<br />
Probabilistic Financial Community Models with Latent Dirichlet Allocation for Financial Supply Chains<br />
Zheng Xu and Louiqa Raschid, University of Maryland<br />
[https://drive.google.com/open?id=0BzTeYQSh4QTsWWJ4b0RiMWFaSlk]<br />
<br />
If you want the gory details of the tools on the IBM System T platform that were developed ...<br />
Exploiting Lists of Names for Named Entity Identification of Financial Institutions from <br />
Unstructured Documents<br />
Zheng Xu (University of Maryland) and Douglas Burdick (IBM) and Louiqa Raschid (University of Maryland)<br />
[http://arxiv.org/abs/1602.04427]<br />
<br />
The dataset is available for research.</div>Louiqahttps://wiki.umiacs.umd.edu/clip/ngfci/index.php/DIBBsDIBBs2014-04-15T22:22:19Z<p>Louiqa: </p>
<hr />
<div> NSF Data Infrastructure Building Blocks<br />
[http://www.nsf.gov/pubs/2014/nsf14530/nsf14530.htm NSF DIBBs CFP]<br />
[[Media:Proposal.pdf]] - Proposal submitted April 2014.</div>Louiqa