Anke Thede

Applicability of Customer Churn Forecasts in a Non-Contractual Setting

As selling a product to an existing customer is much more cost effective than acquiring new customers companies increasingly focus on retaining profitable customers rather than concentrating all marketing actions on the acquisition of new customers. For retaining customers it is very important to be able to predict whether a customer is still active. Effectless marketing expenses directed towards already inactive customers can be avoided and more intensive marketing actions can be taken in order to support active customers’ purchase intentions. Several methods exist that can be used to predict customer activity. In this paper we apply a stochastic and a data mining method to real-life B2B purchase histories and compare the usability and the quality of churn prediction of each of the methods in a non-contractual B2B environment.

Integration of Goods Delivery Supervision into E-commerce Supply Chain

One of the benefits of electronic commerce is the gain of speed and easiness through electronic delivery of commerce data. Only the data specifying the delivery of the physical goods is not integrated into the electronic supply chain and therefore presents an unpleasant interruption of integration. The goal is to find a way to transfer delivery and goods related data electronically to gain rapidity and reliability. This paper focusses on supervisory data of sensitive goods and describes an implementation of a technique to store these data electronically and to integrate their flow seamlessly into the enterprise systems. It also gives an outlook to further utilization and improvements.

Others Also Use: A Robust Recommender System for Scientific Libraries

Scientific digital library systems are a very promising application area for value-added expert advice services. Such systems could significantly reduce the search and evaluation costs of information products for students and scientists. This holds for pure digital libraries as well as for traditional scientific libraries with online public access catalogs (OPAC). In this contribution we first outline different types of recommendation services for scientific libraries and their general integration strategies. Then we focus on a recommender system based on log file analysis that is fully operational within the legacy library system of the Universitat Karlsruhe (TH) since June 2002. Its underlying mathematical model, the implementation within the OPAC, as well as the first user evaluation is presented.

An integration strategy for distributed recommender services in legacy library systems

Scientific library systems are a very promising application area for recommender services. Scientific libraries could easily develop customer-oriented service portals in the style of amazon.com. Students, university teachers and researchers can reduce their transaction cost (i.e. search and evaluation cost of information products). For librarians, the advantage is an improvement of the customer support by recommendations and the additional support in marketing research, product evaluation, and book selection. In this contribution we present a strategy for integrating a behavior-based distributed recommender service in legacy library systems with minimal changes in the legacy systems.

Clustering of Large Document Sets with Restricted Random Walks on Usage Histories

Due to their time complexity, conventional clustering methods often cannot cope with large data sets like bibliographic data in a scientific library. We will present a method for clustering library documents according to usage histories that is based on the exploration of object sets using restricted random walks.

A single sign-on protocol for distributed web applications based on standard internet mechanisms

Personal data has to be entered many times and each user has to memorize different username and password combinations. This reduces system security as users tend to either use passwords that are very easy to guess, or they write them down, or they use the same password for many different accounts. It also increases the cost of the administration of the user accounts. having to login only once. The system is based on standard internet mechanisms. It is composed of different servers that provide authentication and authorization services and is based on cookie technology. The system is designed to be implemented in a heterogenous environment with independent and diverse service providers. The communication between the servers is done via Web services. Additionally, plug-ins are available for other protocols that allow for easy integration of existing authentication and authorization components. A prototype system is operational at the Schroff Stiftungslehrstuhl Information Services and Electronic Markets.

Comparing Simple Association-Rules and Repeat-Buying Based Recommender Systems in a B2B Environment

In this contribution we present a systematic evaluation and comparison of recommender systems based on simple association rules and on repeat-buying theory. Both recommender services are based on the customer purchase histories of a medium-sized B2B-merchant for computer accessories. With the help of product managers an evaluation set for recommendations was generated. With regard to this evaluation set, recommendations produced by both methods are evaluated and several error measures are computed. This provides an empirical test whether frequent item sets or outliers of a stochastic purchase incidence model are suitable concepts for automatically generating recommendations. Furthermore, the loss functions (performance measures) of the two methods are compared and the sensitivity with regard to a misspecification of the model parameters is discussed.

Using Shadow Prices to Reveal Personal Preferences in a Two-Stage Assignment Problem

Process and apparatus for extracting an organic liquid from an organic liquid solute/solvent mixture. The mixture is contacted with a fluid extractant which is at a temperature and pressure to render the extractant a solvent for the solute but not for the solvent. The resulting fluid extract of the solute is then depressurized to give a still feed which is distilled to form still overhead vapors and liquid still bottoms. The enthalpy required to effect this distillation is provided by compressing the still overhead vapors to heat them and indirectly to heat the still feed. The process is particularly suitable for separating mixtures which form azeotropes, e.g., oxygenated hydrocarbon/water mixtures. The energy required in this process is much less than that required to separate such mixtures by conventional distillation techniques.

A Two-Phase Grammar-Based Genetic Algorithm for a Workshop Scheduling Problem

In this contribution we present a two-phase grammar-based genetic algorithm that we use to solve the problem of workshop scheduling in an educational environment which respects partial preferences of participants. The solution respects constraints on workshop capacities and allows for different schedule types. We approach this problem by defining a grammar which defines a language for expressing the restrictions on workshops and participants. A word of this formal language represents a solution which by definition of the language is always feasible. For each feasible schedule the fitness is the result of optimizing the group’s social welfare function which is defined as the sum of the individual utility functions as expressed by the partial preferences. This optimization is achieved with an order based genetic algorithm which assigns to each participant his personal schedule.