Service Analytics

Permanent URI for this collectionhttps://hdl.handle.net/10125/107441

Browse

Now showing 1 - 9 of 9

Data Annotation for Support Ticket Data: A Literature Review
(2024-01-03) Fuchs, Simon; Schnellbach, Janik; Schmidt, Lukas; Wittges, Holger
Supervised Machine Learning is still the most prevalent Machine Learning approach used across the field of Natural Language Processing. As it needs labels to work properly, labeling text data sets is a discerning step in supervised Machine Learning projects. Many industry projects involving supervised Machine Learning never reach a productive phase due to the absence of sufficient labeled data. Against this background, we conducted a Literature Review investigating state of the art approaches to label text data sets for later Natural Language Processing projects. We concentrated on solutions that could be applicable to annotate a support ticket data set. We found that there are three major approaches: Crowdsourcing, Learning Algorithms and Weak Supervision. We also found, that in annotation projects there seems to be an assessment between label quality and cost/effort. We discuss our findings and share our thoughts on the special challenges of annotating a support ticket data set.
Improving Query Performance in RaaS: A Framework for Reputation As a Service in Decentralized Marketplaces
(2024-01-03) Mukkamala, Ravi; Olariu, Stephen; Aljohani, Meshari
With the increasing trend toward online shopping, there is an increasing need to provide reliable reputation management systems that aid buyers and sellers in making intelligent and accurate choices. In addition to the centrally controlled systems such as Amazon.com, several decentralized marketplaces are also coming up. Blockchain technologies and the associated smart contracts are found to be suitable to build reputation management systems in decentralized marketplaces. In this paper, we provide a brief overview of RaaS, a framework to manage Reputation as a Service in decentralized marketplaces. RaaS framework has blockchains and smart contracts as its foundation to guarantee transparency and immutability of reviews. In particular, we address the performance issues faced by such blockchain-based reputation systems, and propose two schemes that address the performance challenges. We provide analytical performance predictions and verify, by extensive simulations, the accuracy of our analytical predictions. Our schemes are shown to result in more efficient query execution with enhanced block structure schemes, with an improvement of up to 81\% in query execution responses when compared to earlier work.
Increasing normal approximation in psychometric health care data analyses using a compositional data approach
(2024-01-03) Lehmann, Rene; Vogt, Bodo
Psychometric health care focuses on the development and improvement of psychotherapeutic measures. Adequate psychological profiling and advanced statistical evaluation are fundamental to assessing the efficacy and measure-associated benefits. Consider psychological constructs operationalized as means or sums of item response values of bipolar Likert scales. Using estimates of the effect size and statistical tests the relevance of a psychotherapeutic measure can be assessed, e.g., via the computation of (partial) correlations of different constructs. Many statistical procedures depend on approximate normal distribution, e.g., t-tests, linear regression and partial least squares path modeling. Increasing the degree of approximate normality of means and sums of item responses enhances the quality of statistical evaluations. Via simulation we provide evidence that applying the isometric log-ratio (ilr) transformation to bipolar Likert scales data prior to the computation of item response means or sums increases the degree of approximate normality. That is, a shift towards normality is observed enhancing the quality of subsequent statistical analyses. As a result, the quality of statistical evaluations enhances. The reliability of psychological diagnostics increases and the development of psychometric scales can be improved enhancing patient welfare. Moreover, reliability and significance affect grant funding in health economics.
OptiGuide: An Efficient Domain-Independent Package Recommender System Based on Multi-Objective Optimization and User Decision Guidance
(2024-01-03) Almanie, Tahani; Brodsky, Alexander
While traditional recommender systems focus on single items, there is an emerging demand for package recommenders that can suggest composite items based on multiple criteria. For instance, they can recommend a combination of dishes based on price, cuisine, and dietary restrictions. Several challenges arise when dealing with package recommenders, including the complexity of the decision-making process and the need of handling trade-offs among conflicting objectives. We introduce OptiGuide, a domain-independent package recommender system that uses efficient multi-objective optimization techniques to guide users effectively in finding their Pareto-optimal recommendations. The user is engaged in the decision-making by capturing their user-system interactions and offering customization options to help them find their optimal recommendation. The system employs preprocessing algorithms to balance the need for quick response times with the computational complexity of the optimization process. A dynamic configuration mechanism is adopted using a pluggable analytic model to enable system versatility across diverse domains.
Sports digitalization – realizing the potential value of tracking technologies in professional sports organizations
(2024-01-03) Konzag, Henrike; Sølvkær Schütz, Nanna
To improve performance on both individual and team levels, professional sports clubs increasingly follow the push for digitalization and adopt digital technologies. While the use of digital technologies is generally associated with great potential, we still know little about the realized value of these technologies. Against this backdrop, we investigated the implementation of a tracking technology in 13 German Handball-Bundesliga clubs to explore both the potential and realized values of such technology. We organized our study as an explorative, multiple case study and collected data by interviewing both clubs’ representatives and the technology vendor. Initial results point to the realized value of tracking technologies deriving from several conversion contingencies. We contribute to the literature on information systems (IS) value creation in the context of professional sports organizations and offer initial empirical insights into the value realization process.
Loss-Leader Pricing Strategies for Personalized Bundles under Customer Choice
(2024-01-03) Xue, Zhengliang; Subramanian, Shiva; Ettl, Markus
This paper considers the pricing of multi-product request-for-quotes (RFQs) that are configured by a buyer based on a large number of products or services offered in a seller’s product catalog. The buyer submits an RFQ for a desired bundle of line items in a bid configuration to a seller. The seller reviews the configuration and offers an approved price for each line item in the bundle. The buyer can selectively purchase any combination of products or services in a bundle configuration at the seller’s approved prices. In addition to a line-item pricing approach, we propose a novel loss-leader model that uses machine learning to calibrate the buyer’s preferences among correlated line items, and dynamically optimizes the prices of any configuration to maximize the seller’s expected profit. The pricing strategies were implemented in a business-to-business (B2B) sales environment with a multinational technology company. Counterfactual analysis shows that loss-leader pricing can generate more than ten percent lift in gross profit over existing pricing practices.
Context-aware Explanations of Accurate Predictions in Service Processes
(2024-01-03) Weinzierl, Sven; Zilker, Sandra; Brunk, Jens; Revoredo, Kate; Matzner, Martin; Becker, Jörg
The performance of a service process can be improved by the early anticipation of future behavior, such as predicting the next activity using predictive business process monitoring (PBPM). Recent PBPM techniques are based on deep neural networks (DNNs) and consider the process context to create accurate predictions. To provide explainability of these predictions, model-agnostic explainable AI (XAI) methods, for example, SHAP, can be used. However, creating these explanations is time-consuming and, therefore, not applicable to service processes where customers are involved. In this paper, we propose a context-aware DNN-based technique to efficiently create meaningful explanations of next activity predictions using layer-wise relevance propagation. We evaluate the predictive quality and the explanation creation time, using three real-life service event logs. Further, we demonstrate its visual output, highlighting its utility for end-users.
Predicting Customer Satisfaction in Service Processes Using Multilingual Large Language Models
(2024-01-03) Liessmann, Annina; Zilker, Sandra; Weinzierl, Sven; Sukhareva, Maria; Matzner, Martin
The huge amount of data recorded during business process executions in today’s organizations creates the need to leverage this data. While most existing business process monitoring methods are capable of including structured context information, the incorporation of unstructured information, for example, text, has rarely been researched. Recent advances in natural language processing offer tools to generate contextualized sentence embeddings, which capture more information in human language than ever before. Especially in service processes, such as the incident management process, a variety of unstructured information is created throughout the execution, representing a relevant use case for incorporating (multilingual) text. To close this research gap, a method exploiting multilingual text for predicting the outcome of a service process is presented. Multilingual large language models are used to generate sentence embeddings for unstructured text created during process execution. After instantiating the method, an evaluation was performed using a real-life event log to predict customer satisfaction in the incident management process at a German multinational corporation. We show that text incorporation improves predictive performance.
Introduction to the Minitrack on Service Analytics
(2024-01-03) Kühl, Niklas; Setzer, Thomas; Fromm, Hansjoerg; VöSsing, Michael

Browse

Recent Submissions