Service Analytics Minitrack

Permanent URI for this collection

Research topics addressed in this minitrack include the applicability of basic and advanced analytics to different service systems, the state-of-the-art of service analytics methodologies and tool-support, and the investigation of benefits resulting from the application of service analytics.

This minitrack will serve as a forum for researchers and practitioners to share progress in the study of these and related themes. Submissions on, but not limited to, the following topics are encouraged:

  • Web Usage Mining and Web Personalization
  • Data Mining
  • Machine Learning applied to Services
  • Recommender Systems for Services
  • Social Network Analytics applied to Services
  • Privacy Issues resulting from Service Analytics
  • Fraud Analytics for Service Systems
  • Analysis and Prediction of User Behavior in Mobile Phone Systems
  • Analysis and Prediction of Driver Behavior in Traffic Situations
  • Analysis and Exploitation of Floating Car Data
  • Electricity Consumption Analysis using Smart Meter Data
  • Analytics for Healthcare Services
  • Analysis and Prediction of IT Service Demand Patterns
  • Analysis of Service Problem Reports
  • Industrial Service Analytics and Optimization
  • Sports Analytics

Minitrack Co-Chairs:

Hansjoerg Fromm (Primary Contact)
Karlsruhe Institute of Technology (KIT)

Gerhard Satzger
IBM Germany

Thomas Setzer
Karlsruhe Institute of Technology (KIT)


Recent Submissions

Now showing 1 - 6 of 6
  • Item
    Smart Data Selection and Reduction for Electric Vehicle Service Analytics
    ( 2017-01-04) Schoch, Jennifer ; Staudt, Philipp ; Setzer, Thomas
    Battery electric vehicles (BEV) are increasingly used in mobility services such as car-sharing. A severe problem with BEV is battery degradation, leading to a reduction of the already very limited range of a BEV. Analytic models are required to determine the impact of service usage to provide guidance on how to drive and charge and also to support service tasks such as predictive maintenance. However, while the increasing number of sensor data in automotive applications allows for more fine-grained model parameterization and better predictive outcomes, in practical settings the amount of storage and transmission bandwidth is limited by technical and economical considerations. By means of a simulation-based analysis, dynamic user behavior is simulated based on real-world driving profiles parameterized by different driver characteristics and ambient conditions. We find that by using a shrinked subset of variables the required storage can be reduced considerably at low costs in terms of only slightly decreased predictive accuracy. \
  • Item
    Road Condition Estimation Based on Heterogeneous Extended Floating Car Data
    ( 2017-01-04) Laubis, Kevin ; Simko, Viliam ; Schuller, Alexander ; Weinhardt, Christof
    Road condition estimation based on Extended Floating Car Data (XFCD) from smart devices allows for determining given quality indicators like the international roughness index (IRI). Such approaches currently face the challenge to utilize measurements from heterogeneous sources. This paper investigates how a statistical learning based self-calibration overcomes individual sensor characteristics. We investigate how well the approach handles variations in the sensing frequency. Since the self-calibration approach requires the training of individual models for each participant, it is examined how a reduction of the amount of data sent to the backend system for training purposes affects the model performance. We show that reducing the amount of data by approximately 50 % does not reduce the models’ performance. Likewise, we observe that the approach can handle sensing frequencies up to 25 Hz without a performance reduction compared to the baseline scenario with 50 Hz.
  • Item
    Parameter Tolerance in Capacity Planning Models
    ( 2017-01-04) Leung, Ying Tat ; Kamath, Manjunath ; Ma, Juan
    In capacity planning for a service operation, analytical models based on queueing theory allow the user to quickly estimate the capacity required and to easily experiment with different system designs or configurations, for a given set of input parameters. An input parameter of the model could be inaccurate or may not be known beyond a good guess. In order to determine if the analysis results (and hence the system design) are robust to parameter estimation errors, sensitivity analysis can be performed. We study an alternative approach that involves specifying a tolerance range of a system performance measure and calculating a feasible region of the uncertain parameters for which the performance measure will be within the tolerance range. We illustrate this approach using basic exponential queueing models as well as a model of an order fulfillment operation in a distribution center.
  • Item
    GreedyBoost: An Accurate, Efficient and Flexible Ensemble Method for B2B Recommendations
    ( 2017-01-04) Zhang, Weipeng ; Enders, Tobias ; Li, Dongsheng
    Recommender systems have achieved great success in finding relevant products and services for individual customers, e.g. in B2C markets, during recent years. \ However, due to the diversity of enterprise clients' requirements it is still an open question on how to successfully apply existing recommendation techniques in the B2B domain. \ \ This paper presents GreedyBoost --- an accurate, efficient and flexible ensemble method for product and service recommendations in the B2B domain. Given a set of base models, GreedyBoost can sequentially add base models to the ensemble by a linear approach to minimize training error, so that the ensemble process is efficient. Meanwhile, GreedyBoost does not have any special requirement on base models and evaluation metrics, so that any kind of client requirements and sale \\& distribution purposes can be adapted. Experimental results on real-world B2B data demonstrate that GreedyBoost can achieve higher recommendation accuracy compared with two popular ensemble methods.
  • Item
    Generating Rental Data for Car Sharing Relocation Simulations on the Example of Station-Based One-Way Car Sharing
    ( 2017-01-04) Brendel, Alfred Benedikt ; Rockenkamm, Christian ; Kolbe, Lutz Maria
    Developing sophisticated car sharing simulations is a major task to improve car sharing as a sustainable means of transportation, because new \ algorithms for enhancing car sharing efficiency are formulated using them. \ \ Simulations rely on input data, which is often gathered in car sharing systems or artificially generated. Real-world data is often incomplete and biased while artificial data is mostly generated based on initial assumptions. Therefore, developing new ways for generating testing data is an important task for future research. \ \ In this paper, we propose a new approach for generating car sharing data for relocation simulations by utilizing machine learning. Based on real-world data, we could show that a combined methods approach consisting of a Gaussian Mixture Model and two classification trees can generate appropriate artificial testing data.
  • Item
    Introduction to Service Analytics Minitrack
    ( 2017-01-04) Fromm, Hansjoerg ; Satzger, Gerhard ; Setzer, Thomas