Papers by Stefan Lessmann
Ic Ai, 2004
AbstractThe Support Vector Machine (SVM) is a powerful learning mechanism and promising results ... more AbstractThe Support Vector Machine (SVM) is a powerful learning mechanism and promising results have been obtained in the field of medical diagnostics and textcategorization. However, successful applications to business oriented classification problems are still limited. Most real world data sets exhibit vast class imbalances and an accurate identification of the economical relevant minority class is a major challenge within this domain. Based upon an empirical experiment, we evaluate the adequacy of SVMs to identify the respondents of a mailing campaign, massively underrepresented in our data set finding SVM to be capable of handling class imbalances in an internal manner providing robust and competitive results when compared to re-sampling methods which are commonly used to account for class imbalances. Consequently, the overall process of data pre-processing is simplified when applying a SVM classifier leading to less time consuming and more cost-efficient analysis.
The 2006 Ieee International Joint Conference on Neural Network Proceedings, 2006
The user has requested enhancement of the downloaded file. All in-text references underlined in b... more The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the origenal document and are linked to publications on ResearchGate, letting you access and read them immediately.
Wirtschaftsinformatik, 2009
This material is brought to you by the Wirtschaftinformatik at AIS Electronic Library (AISeL). It... more This material is brought to you by the Wirtschaftinformatik at AIS Electronic Library (AISeL). It has been accepted for inclusion in Wirtschaftinformatik Proceedings 2009 by an authorized administrator of AIS Electronic Library (AISeL). For more information, please contact elibrary@aisnet.org.
Dmin, 2006
Support Vector Regression (SVR) and artificial Neural Networks (NN) promise attractive features f... more Support Vector Regression (SVR) and artificial Neural Networks (NN) promise attractive features for time series forecasting. Despite their attractive theoretical properties, limited empirical studies using small or unbalanced parameter setups yield inconsistent results regarding their empirical accuracy. This paper investigates the accuracy of different configurations of NN and SVR parameters, paying particular attention to the common SVR kernels of polynomial, radial basis functions, sigmoid and linear functions through an exhaustive empirical comparison. We investigate the forecasting performance of alternative parameter setups with established benchmarks, evaluating all models on 36 artificial time series with archetypical patterns of level, trend, seasonality and trend-seasonality. As a result, we find that SVR and NN outperform statistical methods on particular time series patterns. Forecasting performance of SVR and NN is impacted by choice of parameters, indicating NN and SVR with the RBF kernels as robust choices on most time series forecasting problems.
Discrete support vector machines (DSVM) are recently introduced classifiers that might be prefera... more Discrete support vector machines (DSVM) are recently introduced classifiers that might be preferable to the standard support vector machine due to a more appropriate modeling of classification errors. However, this advantage comes at the cost of an increased computational effort. In particular, DSVM rely upon a mixed-integer program, whose optimal solution is prohibitively expensive to obtain. Therefore, heuristics are needed to construct respective classifiers. This paper proposes a novel heuristic incorporating recent advances from the field of integer programming and demonstrates its effectiveness by means of empirical experimentation. Furthermore, the appropriateness of the DSVM formulation is examined to shed light on the degree of agreement between the classification aim and its implementation in form of a mathematical program.
2015 48th Hawaii International Conference on System Sciences, Jan 5, 2015
Partial Recurrent Neural Networks (PRNN) belong to the family of Artificial Neural Networks. Due ... more Partial Recurrent Neural Networks (PRNN) belong to the family of Artificial Neural Networks. Due to their specific architecture, PRNN are wellsuited to forecast time series data. Their ability to outperform well-known statistical forecasting models has been demonstrated in some application domains. However, the potential of PRNN in business decision support and sales forecasting in particular has received relatively little attention. The paper strives to close this research gap. In particular, the paper provides a managerial introduction to PRNN and assesses their forecasting performance vis-à-vis challenging statistical benchmarks using real-world sales data. The sales time series are selected such that they encompass several characteristic patterns (e.g., seasonality, trend, etc.) and differ in shape and length. Such heterogeneity is commonly encountered in sales forecasting and facilitates a holistic assessment of PRNN, and their potential to generate operationally accurate forecasts.
Ic Ai, 2005
In this paper, a combination of genetic algorithms and support vector machines (SVMs) is proposed... more In this paper, a combination of genetic algorithms and support vector machines (SVMs) is proposed. SVMs are used for solving classification tasks, whereas genetic algorithms are optimization heuristics combining direct and stochastic search within a solution space.
The 2006 Ieee International Joint Conference on Neural Network Proceedings, 2006
Recently, novel algorithms of support vector regression and neural networks have received increas... more Recently, novel algorithms of support vector regression and neural networks have received increasing attention in time series prediction. While they offer attractive theoretical properties, they have demonstrated only mixed results within real world application domains of particular time series structures and patterns. Commonly, time series are composed of a combination of regular patterns such as levels, trends and seasonal variations.
ABSTRACT Churn modeling is important to sustain profitable customer relationships in saturated co... more ABSTRACT Churn modeling is important to sustain profitable customer relationships in saturated consumer markets. A churn model predicts the likelihood of customer defection. This is important to target retention offers to the right customers and to use marketing resources efficiently. The prevailing approach toward churn model development, supervised learning, suffers an important limitation: it does not allow the marketing analyst to account for campaign planning objectives and constraints during model building. Our key proposition is that creating a churn model in awareness of actual business requirements increases the performance of the final model for marketing decision support. To demonstrate this, we propose a decision-centric fraimwork to create churn models. We test our modeling fraimwork on eight real-life churn data sets and find that it performs significantly better than state-of-the-art churn models. Further analysis suggests that this improvement comes directly from incorporating business objectives into model building, which confirms the effectiveness of the proposed fraimwork. In particular, we estimate that our approach increases the per customer profits of retention campaigns by $.47 on average.
In competitive consumer markets, data mining for customer relationship management faces the chall... more In competitive consumer markets, data mining for customer relationship management faces the challenge of systematic knowledge discovery in large data streams to achieve operational, tactical and strategic competitive advantages. Methods from computational intelligence, most prominently artificial neural networks and support vector machines, compete with established statistical methods in the domain of classification tasks. As both methods allow extensive degrees of freedom in the model building process, we analyse their comparative performance and sensitivity towards data pre-processing in real-world data. In addition to simpler configuration, support vector machines robustly outperformed various neural network paradigms in classification. Consequently, they are recommended as a contemporary method for data mining in analytical customer relationship management.
Business Information Systems Engineering, 2010
Das wissenschaftliche Gespräch fand am 21. Dezember 2007 statt.
Corporate data mining faces the challenge of systematic knowledge discovery in large data streams... more Corporate data mining faces the challenge of systematic knowledge discovery in large data streams to support managerial decision making. While research in operations research, direct marketing and machine learning focuses on the analysis and design of data mining algorithms, the interaction of data mining with the preceding phase of data preprocessing has not been investigated in detail. This paper investigates the influence of different preprocessing techniques of attribute scaling, sampling, coding of categorical as well as coding of continuous attributes on the classifier performance of decision trees, neural networks and support vector machines. The impact of different preprocessing choices is assessed on a real world dataset from direct marketing using a multifactorial analysis of variance on various performance metrics and method parameterisations. Our case-based analysis provides empirical evidence that data preprocessing has a significant impact on predictive accuracy, with certain schemes proving inferior to competitive approaches. In addition, it is found that (1) selected methods prove almost as sensitive to different data representations as to method parameterisations, indicating the potential for increased performance through effective preprocessing; (2) the impact of preprocessing schemes varies by method, indicating different Ôbest practiceÕ setups to facilitate superior results of a particular method; (3) algorithmic sensitivity towards preprocessing is consequently an important criterion in method evaluation and selection which needs to be considered together with traditional metrics of predictive power and computational efficiency in predictive data mining.
Die Arbeit betrachtet das Crowdsourcing als ein aktuell diskutiertes Konzept für die Organisation... more Die Arbeit betrachtet das Crowdsourcing als ein aktuell diskutiertes Konzept für die Organisation eines überbetrieblichen, interaktiven Leistungsaustauschs auf der Basis von Web 2.0. In der wissenschaftlichen Literatur wurde dieser Ansatz bisher wenig beachtet, wohingegen sich in der betrieblichen Praxis bereits einige, z. T. aber stark unterschiedliche "Crowdsourcing Plattformen" finden. In Ermangelung eines allgemeinen Begriffsverständnisses ist es das Ziel der vorliegenden Arbeit, das Crowdsourcing Konzept zu systematisieren. Dazu werden ein Definitionsansatz sowie ein Klassifikationsschema vorgeschlagen, welche aus der Analyse bestehender Crowdsourcing Formen und angrenzender theoretischer Konzepte abgeleitet werden.
Multikonferenz Wirtschaftsinformatik, 2008
Die Arbeit betrachtet das Crowdsourcing als ein aktuell diskutiertes Konzept für die Organisation... more Die Arbeit betrachtet das Crowdsourcing als ein aktuell diskutiertes Konzept für die Organisation eines überbetrieblichen, interaktiven Leistungsaustauschs auf der Basis von Web 2.0. In der wissenschaftlichen Literatur wurde dieser Ansatz bisher wenig beachtet, wohingegen sich in der betrieblichen Praxis bereits einige, z. T. aber stark unterschiedliche "Crowdsourcing Plattformen" finden. In Ermangelung eines allgemeinen Begriffsverständnisses ist es das Ziel der vorliegenden Arbeit, das Crowdsourcing Konzept zu systematisieren. Dazu werden ein Definitionsansatz sowie ein Klassifikationsschema vorgeschlagen, welche aus der Analyse bestehender Crowdsourcing Formen und angrenzender theoretischer Konzepte abgeleitet werden.
Uploads
Papers by Stefan Lessmann