Filters








1,387 Hits in 9.0 sec

A Novel Example-Dependent Cost-Sensitive Stacking Classifier to Identify Tax Return Defaulters

Sanat Bhargava, M. Ravi Kumar, Priya Mehta, Jithin Mathews, Sandeep Kumar, Ch Sobhan Babu
2021 Business Information Systems  
This paper proposes a framework for an example-dependent cost-sensitive stacking classifier that uses cost-insensitive classifiers as base generalizers to make predictions on the input space.  ...  Additionally, we develop an empirical study showing that our approach is more adept at identifying potential tax return defaulters than existing example-dependent cost-sensitive classification algorithms  ...  To alleviate this problem, we propose a novel framework for example-dependent cost-sensitive stacked classifiers that give a competitive model performance and increased savings compared to cost-insensitive  ... 
doi:10.52825/bis.v1i.61 fatcat:3zsm7wfd4vemzemlornjgswryu

Can Deep Learning Predict Risky Retail Investors? A Case Study in Financial Risk Behavior Forecasting [article]

Yaodong Yang, Alisa Kolesnikova, Stefan Lessmann, Tiejun Ma, Ming-Chien Sung, Johnnie E.V. Johnson
2019 arXiv   pre-print
The paper examines the potential of deep learning to support decisions in financial risk management.  ...  Deep learning promises a remedy.  ...  Teunter, for his efforts in handling our paper and are thankful to three anonymous reviewers whose feedback has helped tremendously to improve earlier versions of the paper.  ... 
arXiv:1812.06175v3 fatcat:aswsbfb4xrbozellaic7fireom

Ensemble Model of the Financial Distress Prediction in Visegrad Group Countries

Michal Pavlicko, Marek Durica, Jaroslav Mazanec
2021 Mathematics  
Five variables were selected as significant predictors in the model: current ratio, return on equity, return on assets, debt ratio, and net working capital.  ...  Moreover, the ensemble model is a new technique in the Visegrad Group (V4) compared with other prediction models.  ...  The results identify non-default and default companies. If the default probability is more than 50%, the company is a non-prosperous entity.  ... 
doi:10.3390/math9161886 fatcat:zmwih2vzzzgsnlvcjgy3ccl62a

One-class classification: taxonomy of study and review of techniques

Shehroz S. Khan, Michael G. Madden
2014 Knowledge engineering review (Print)  
This unique situation constrains the learning of efficient classifiers by defining class boundary just with the knowledge of positive class.  ...  One-class classification (OCC) algorithms aim to build classification models when the negative class is either absent, poorly sampled or not well defined.  ...  The value of the threshold (default 1.0) can be changed to either higher or lower values to change the detection sensitivity of the classifier.  ... 
doi:10.1017/s026988891300043x fatcat:djdcvpij7jhs7gygtrtihq3dia

Machine Learning with a Reject Option: A survey [article]

Kilian Hendrickx, Lorenzo Perini, Dries Van der Plas, Wannes Meert, Jesse Davis
2021 arXiv   pre-print
Finally, we provide examples of relevant application domains and show how machine learning with rejection relates to other machine learning research areas.  ...  This machine learning subfield enables machine learning models to abstain from making a prediction when likely to make a mistake.  ...  Finally, a trivial approach is to simply reject a default percentage of examples regardless of the testing dataset (Tax and Duin 2008; Denis and Hebiri 2020) .  ... 
arXiv:2107.11277v1 fatcat:hsatf7zp7re4vo52iyrbcllbzu

Investment: Fundamentals and Finance

Simon Gilchrist
1998 NBER macroeconomics annual  
+s ,B (+s+1 }=o =0 s =i \fc=i / subject to 16. For example, Moyen (1997) derives an equilibrium debt premium generated by default costs. 17.  ...  In a forward-looking model, investment depends on marginal Q, the present value of expected future marginal returns to capital.  ... 
doi:10.1086/ma.13.4623744 fatcat:wejsq2cam5d2hn6naiuvlulny4

Hunter in the Dark: Discover Anomalous Network Activity Using Deep Ensemble Network [article]

Shiyi Yang, Hui Guo, Nour Moustafa
2021 arXiv   pre-print
However, existing designs achieve a high threat detection performance at the cost of a large number of false alarms, leading to alert fatigue.  ...  It uses a deep ensemble network (trained through supervised learning) to detect anomalous network activities and exploits an unsupervised learning-based scheme to trim off mis-detection results.  ...  While ML-based IDSs have a high capability of novel threat perception, existing designs achieve a high attack detection performance often at the cost of a large number of false alarms.  ... 
arXiv:2105.09157v4 fatcat:zkcteve4l5ezhfg5xetq5nku2y

Local Bank Financial Constraints and Firm Access to External Finance

DANIEL PARAVISINI
2008 Journal of Finance  
Using novel data to measure risk and return on marginal lending, I show that the profitability of loans does not decline during lending expansions.  ...  Using novel data to measure risk and return on marginal lending, I show that the profitability of loans does not decline during lending expansions.  ...  For example, an increase in deposits might decrease the risk profile of the bank and reduce the cost of capital, or signal better future lending prospects and a higher marginal returns on lending.  ... 
doi:10.1111/j.1540-6261.2008.01393.x fatcat:4qry6lw6znffrg7td34o5g5s4e

News-based trading strategies

Stefan Feuerriegel, Helmut Prendinger
2016 Decision Support Systems  
In this paper, we design trading strategies that utilize textual news in order to obtain profits on the basis of novel information entering the market.  ...  Interestingly, this so-called news sentiment works fairly well in explaining stock returns.  ...  These are then passed into a neural network to classify future performances.  ... 
doi:10.1016/j.dss.2016.06.020 fatcat:dw55lpgpebghrg5kyervzb4ljy

THE AMERICAN MUSEUM OF NATURAL HISTORY

1927 Science  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1126/science.65.1672.31 fatcat:o6e4olz4qjbxjd5gc3daqoyzgm

THE AMERICAN MUSEUM OF NATURAL HISTORY

1911 Science  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1126/science.33.839.142 pmid:17731786 fatcat:rktmdddq3jblpniceihkfymqji

THE AMERICAN MUSEUM OF NATURAL HISTORY

1903 Science  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1126/science.17.439.874 pmid:17740807 fatcat:gq6kbw4cxffsrhdkwlap27zqga

The American Museum of Natural History

1870 Scientific American  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1038/scientificamerican05141870-319 fatcat:oxbmiiy2x5gj5botnxdsoqo6om

American Museum of Natural History

Henry Gee
1994 Nature  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1038/369087b0 fatcat:daq5fif7fzg7lmrami6etrxco4

The American Museum of Natural History

1922 Nature  
These commands switch jobs when a novel processor appears 20% faster than the one currently used. Success in phylogenetic analysis has become increasingly dependent on computational speed.  ...  Dynamic programming (as in matrix characters above) is used to identify the best HTU sequences and determine cladogram cost.  ...  Default = none autapomorphy An apomorphy unique to one taxon.  ... 
doi:10.1038/110880a0 fatcat:x2q5eso3bfempe7kike2ang5qa
« Previous Showing results 1 — 15 out of 1,387 results