Deron Liang

The effect of feature selection on financial distress prediction

Financial distress prediction is always important for financial institutions in order for them to assess the financial health of enterprises and individuals. Bankruptcy prediction and credit scoring are two important issues in financial distress prediction where various statistical and machine learning techniques have been employed to develop financial prediction models. Since there are no generally agreed upon financial ratios as input features for model development, many studies consider feature selection as a pre-processing step in data mining before constructing the models. However, most works only focused on applying specific feature selection methods over either bankruptcy prediction or credit scoring problem domains. In this work, a comprehensive study is conducted to examine the effect of performing filter and wrapper based feature selection methods on financial distress prediction. In addition, the effect of feature selection on the prediction models obtained using various classification techniques is also investigated. In the experiments, two bankruptcy and two credit datasets are used. In addition, three filter and two wrapper based feature selection methods combined with six different prediction models are studied. Our experimental results show that there is no the best combination of the feature selection method and the classification technique over the four datasets. Moreover, depending on the chosen techniques, performing feature selection does not always improve the prediction performance. However, on average performing the genetic algorithm and logistic regression for feature selection can provide prediction improvements over the credit and bankruptcy datasets respectively.

Financial ratios and corporate governance indicators in bankruptcy prediction: A comprehensive study

Effective bankruptcy prediction is critical for financial institutions to make appropriate lending decisions. In general, the input variables (or features), such as financial ratios, and prediction techniques, such as statistical and machine learning techniques, are the two most important factors affecting the prediction performance. While many related works have proposed novel prediction techniques, very few have analyzed the discriminatory power of the features related to bankruptcy prediction. In the literature, in addition to financial ratios (FRs), corporate governance indicators (CGIs) have been found to be another important type of input variable. However, the prediction performance obtained by combining CGIs and FRs has not been fully examined. Only some selected CGIs and FRs have been used in related studies and the chosen features may differ from study to study. Therefore, the aim of this paper is to assess the prediction performance obtained by combining seven different categories of FRs and five different categories of CGIs. The experimental results, based on a real-world dataset from Taiwan, show that the FR categories of solvency and profitability and the CGI categories of board structure and ownership structure are the most important features in bankruptcy prediction. Specifically, the best prediction model performance is obtained with a combination in terms of prediction accuracy, Type I/II errors, ROC curve, and misclassification cost. However, these findings may not be applicable in some markets where the definition of distressed companies is unclear and the characteristics of corporate governance indicators are not obvious, such as in the Chinese market.

A New Non-intrusive Authentication Approach for Data Protection Based on Mouse Dynamics

Mouse-dynamics-related schemes have been shown to be feasible for user authentication systems, however, the existing approaches are either intrusive or not prompt response. Preventing unauthorized accesses to critical digital assets, namely, data stored in the file management system, is one of the major objectives of user authentication. We therefore propose a non-intrusive approach capable of verifying a user having performed a few times of file-related operations via a mouse. To evaluate the effectiveness of the proposed approach, the mouse movement of the file-related operations in Explorer, which is the most common way to search, open, save, copy, and/or delete files in Windows environments, is used for authentication. The experimental results show that the proposed approach is feasible and has three advantages: 1) it is non-intrusive, 2) it authenticates users in a short period of time, and 3) the quantity of mouse dynamics used for authentication purpose is lightweight.

An Analysis of Using State of the Art Technologies to Implement Real-Time Continuous Assurance

With the integrity of the information in financial reports being questioned and the shift towards more rapid financial reporting, the auditing profession has found that Continuous Assurance is an effective means of facilitating early detection of fraudulent financial reports. However, according to recent surveys, Continuous Assurance has not been widely applied to date. This fact motivates us to investigate if state-of-the-art IT technologies are capable of supporting Continuous Assurance. The contribution of this study is threefold. First, we develop an ISO/IEC 9126-based Continuous Assurance evaluation framework with six technical criteria. Second, based on the proposed framework, we review two (real-time) IT technologies, namely the Embedded Audit Module (EAM) and the Interceptor mechanism, and explore the feasibility of using them to implement real-time Continuous Assurance (CA). Overall, the interceptor approach outperforms the EAM approach, although neither approach satisfies all of the framework’s technical criteria. Third, we find that using the interceptor mechanism in the middleware layer, rather than in other layers, improves the implementation of a real-time auditing interceptor. In light of the proposed evaluation framework, we consider the future development of a middleware interceptor technology that can be used to firmly establish a real-time Continuous Assurance framework.

A Novel Non-intrusive User Authentication Method Based on Touchscreen of Smartphones

In recent years, the functionality of smartphones has been rapidly improved, then, the smartphones are not only used for telecommunication but also for various applications, such as email and social network accessing. These applications raise new security issues to smartphone users, however, the current protection mechanisms of smartphones are not sufficient due to convenience issue and shoulder-surfing issue. We therefore propose a non-intrusive authentication approach based on touch screen of smartphones. To the best of our knowledge, this work is the first publicly reported study that adopts the histogram features of touch screen to build an authentication model for smartphone users. Our empirical results for fifty-five participants show that the proposed approach is feasible. The performance of the proposed approach could be increased if users continuously operate their smartphone after a period of time. Finally, we further discuss the applications and limitations of the proposed approach.

A novel classifier ensemble approach for financial distress prediction

Financial distress prediction is very important to financial institutions who must be able to make critical decisions regarding customer loans. Bankruptcy prediction and credit scoring are the two main aspects considered in financial distress prediction. To assist in this determination, thereby lowering the risk borne by the financial institution, it is necessary to develop effective prediction models for prediction of the likelihood of bankruptcy and estimation of credit risk. A number of financial distress prediction models have been constructed, which utilize various machine learning techniques, such as single classifiers and classifier ensembles, but improving the prediction accuracy is the major research issue. In addition, aside from improving the prediction accuracy, there have been very few studies that specifically consider lowering the Type I error. In practice, Type I errors need to receive careful consideration during model construction because they can affect the cost to the financial institution. In this study, we introduce a classifier ensemble approach designed to reduce the misclassification cost. The outputs produced by multiple classifiers are combined by utilizing the unanimous voting (UV) method to find the final prediction result. Experimental results obtained based on four relevant datasets show that our UV ensemble approach outperforms the baseline single classifiers and classifier ensembles. Specifically, the UV ensemble not only provides relatively good prediction accuracy and minimizes Type I/II errors, but also produces the smallest misclassification cost.

A novel driver identification method using wearables

Today, vehicles have been an essential part of our daily life. According to survey, one-third drivers admit they have left their vehicle while it running, which makes the vehicle an easy target for theft. In this paper, we propose a novel driver identification method with wearable device. This approach can be used in continuous authentication to offer a greater degree of multilevel protection. Our proposed method is based on the hypothesis that a driver has a specific habit to drive a vehicle; and such behavioral biometrics can be captured from the wearable device. In order to validating this hypothesis, we have used the wireless sensor modules to collect drivers behavioral information. Then, the experimental results indicate that the proposed approach is feasible.

The Application of Corporate Governance Indicators With XBRL Technology to Financial Crisis Prediction

ABSTRACT The widespread adoption of eXtensible Business Reporting Language (XBRL) suggests that intelligent software agents can now use financial information disseminated on the Web with high accuracy. Financial data have been widely used by researchers to predict financial crises; however, few studies have considered corporate governance indicators in building prediction models. This article presents a financial crisis prediction model that involves using a genetic algorithm for determining the optimal feature set and support vector machines (SVMs) to be used with XBRL. The experimental results show that the proposed model outperforms models based on only one type of information, either financial or corporate governance. Compared with conventional statistical methods, the proposed SVM model forecasts financial crises more accurately.

A portable interceptor mechanism for SOAP frameworks

An interceptor is a generic architecture pattern, and has been used to resolve specific issues in a number of application domains. Many standard platforms such as CORBA also provide interception interfaces so that an interceptor developed for a specific application can become portable across systems running on the same platform. SOAP frameworks are commonly used platforms to build Web Services. However, there is no standard way to build interceptors portable across current SOAP frameworks, although, some of them provide proprietary interceptor solution within individual framework, such as Axis, XFire, and etc. In this paper, we propose the portable interceptor mechanism (PIM) consisting of a set of application programming interfaces (API) on SOAP engine, a core component of a SOAP framework. An interceptor is able to receive messages passing through the SOAP framework from the SOAP engine via these APIs. Furthermore, the proposed PIM facilitates run-time lifecycle management of interceptors that is a crucial feature to many application domains but is not fully supported by CORBA standard. For concept proving, we implement the proposed PIM on two popular SOAP frameworks, namely, Axis and XFire. We also discuss a number of implementation issues including the performance and reliability of PIM.

A Preliminary Study on Non-Intrusive User Authentication Method Using Smartphone Sensors

This paper proposes a non-intrusive authentication method based on two sensitive apparatus of smartphones, namely, the orientation sensor and the touchscreen. We have found that these two sensors are capable of capturing behavioral biometrics of a user while the user is engaged in relatively stationary activities. The experimental results with respect to two types of flick operating have an equal error rate of about 3.5% and 5%, respectively. To the best of our knowledge, this work is the first publicly reported study that simultaneously adopts the orientation sensor and the touchscreen to build an authentication model for smartphone users. Finally, we show that the proposed approach can be used together with existing intrusive mechanisms, such as password and/or fingerprints, to build a more robust authentication framework for smartphone users.