Detecting stock-price manipulation in an emerging market: The case of Turkey

Authors:
Hulisi Öğüt;M. Mete Doğanay;Ramazan Aktaş
Affiliations:
Department of Business Administration, TOBB University of Economics and Technology, Söğütözü Caddesi No: 43, Ankara 06560, Turkey;Department of Business Administration, Çankaya University, Öğretmenler Caddesi No: 14, Balgat, Ankara 06530, Turkey;Department of Business Administration, TOBB University of Economics and Technology, Söğütözü Caddesi No: 43, Ankara 06560, Turkey
Venue:
Expert Systems with Applications: An International Journal
Year:
2009

Citing 8
Cited 0

The nature of statistical learning theory

The nature of statistical learning theory
A Tutorial on Support Vector Machines for Pattern Recognition

Data Mining and Knowledge Discovery
Asymptotic behaviors of support vector machines with Gaussian kernel

Neural Computation
Churn prediction in subscription services: An application of support vector machines while comparing two parameter-selection techniques

Expert Systems with Applications: An International Journal
Prediction in Marketing Using the Support Vector Machine

Marketing Science
Collusion set detection using graph clustering

Data Mining and Knowledge Discovery
Bankruptcy prediction using support vector machine with optimal choice of kernel function parameters

Expert Systems with Applications: An International Journal
The evaluation of consumer loans using support vector machines

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	12.05

Visualization

Abstract

This paper aims to develop methods that are capable of detecting manipulation in the Istanbul Stock Exchange. We take the difference between manipulated stock's and index's average daily return, average daily change in trading volume and average daily volatility and used these statistics as explanatory variables. The data in post-manipulation and pre-manipulation periods are used as non-manipulated instances while the data in the manipulation period are used as manipulated instances. Test performance of classification accuracy, sensitivity and specificity statistics for Artificial Neural Networks (ANN) and Support Vector Machine (SVM) are compared with the results of discriminant analysis and logistics regression (logit). We found that the data mining techniques (ANN and SVM) are better suited to detect stock-price manipulation than multivariate statistical techniques (discriminant analysis, logistics regression) as the performances of the data mining techniques in terms of total classification accuracy and sensitivity statistics are better than those of multivariate techniques. We also found that unit change in difference between average daily return of manipulated stock and the index has the largest effect while unit change in difference between average daily change in trading volume of manipulated stock and index has the least effect on multivariate classifiers' decision functions.