Dominic Girardi

Interactive knowledge discovery with the doctor-in-the-loop: a practical example of cerebral aneurysms research

Established process models for knowledge discovery find the domain-expert in a customer-like and supervising role. In the field of biomedical research, it is necessary to move the domain-experts into the center of this process with far-reaching consequences for both their research output and the process itself. In this paper, we revise the established process models for knowledge discovery and propose a new process model for domain-expert-driven interactive knowledge discovery. Furthermore, we present a research infrastructure which is adapted to this new process model and demonstrate how the domain-expert can be deeply integrated even into the highly complex data-mining process and data-exploration tasks. We evaluated this approach in the medical domain for the case of cerebral aneurysms research.

Servicing Individual Product Variants within Value Chains with an Ontology

Today companies have to satisfy manifold and changing customer demands. However, the question why and under what circumstances certain product variants cause performance problems within value chain operations remains unsolved. Existing process descriptions, data-models and IT-systems are remarkably supporting planning and optimization efforts, but show significant deficits regarding the integration of heterogeneous internal and cross-company systems. Relevant issues such as the evaluation of demand variety impacts on the subsequent value-adding steps in the value chain are not sufficiently solved. Ontological modelling could notably advance the information exchange in complex value chains and thus enhance value chain flexibility through a semantic harmonization that enables faster and faultless information flows between companies. Further advantages are the distinct reusability, modifiability, extendibility and shareability of ontology-based value chain models and software. Despite a distinct need to advance cross-company integration, ontologies are used in supply chain management only occasionally. The need for methodical enhancement is high. Thus, the objective of this paper is to investigate obtainable benefits of ontological modelling for supply chain management (SCM) – here substantiated by means of a methodological support to improve product variety management in value networks. We provide a literature review, a conceptual framework and an implementation-guide for use in scenarios that represent the problems mentioned. The project-design was developed in an explorative feasibility study based on the sample case of an Austrian manufacturer. An achieved managerial insight that extends current SC-ontology contributions is a conceivable approach on how to gain a proof of concept for cross-company ontology application in value chains.

Ontology-Guided Principal Component Analysis: Reaching the Limits of the Doctor-in-the-Loop

Biomedical research requires deep domain expertise to perform analyses of complex data sets, assisted by mathematical expertise provided by data scientists who design and develop sophisticated methods and tools. Such methods and tools not only require preprocessing of the data, but most of all a meaningful input selection. Usually, data scientists do not have sufficient background knowledge about the origin of the data and the biomedical problems to be solved, consequently a doctor-in-the-loop can be of great help here. In this paper we revise the viability of integrating an analysis guided visualization component in an ontology-guided data infrastructure, exemplified by the principal component analysis. We evaluated this approach by examining the potential for intelligent support of medical experts on the case of cerebral aneurysms research.

Using concept hierarchies to improve calculation of patient similarity

OBJECTIVE We introduce a new distance measure that is better suited than traditional methods at detecting similarities in patient records by referring to a concept hierarchy. MATERIALS AND METHODS The new distance measure improves on distance measures for categorical values by taking the path distance between concepts in a hierarchy into account. We evaluate and compare the new measure on a data set of 836 patients. RESULTS The new measure shows marked improvements over the standard measures, both qualitatively and quantitatively. Using the new measure for clustering patient data reveals structure that is otherwise not visible. Statistical comparisons of distances within patient groups with similar diagnoses shows that the new measure is significantly better at detecting these similarities than the standard measures. CONCLUSION The new distance measure is an improvement over the current standard whenever a hierarchical arrangement of categorical values is available.

A Domain-Expert Centered Process Model for Knowledge Discovery in Medical Research: Putting the Expert-in-the-Loop

Established process models for knowledge discovery see the domain expert in a customer-like, supervising role. In the field of bio-medical research, it is necessary for the domain experts to move into the center of this process with far-reaching consequences for their research work but also for the process itself. We revise the established process models for knowledge discovery and propose a new process model for domain-expert driven knowledge discovery. Furthermore, we present a research infrastructure which is adapted to this new process model and show how the domain expert can be deeply integrated even into the highly complex data mining and machine learning tasks.

A Meta-model Guided Expression Engine

Data acquisition and handling is known to be one of the most severe technical barriers in bio-medical research. In order to counter this problem, we created a generic data acquisition and managing system which can be set up for the given domain of application without the need for programming- or database-skills. The user definitions of the domain data structures are stored into an abstract meta data-model and allow the automatic creation of data-input and -managing interfaces. In order to enable the user to define complex search queries on the data or derive new data out of already existing, a meta-model-guided expression engine was developed. Grammatical and structural meta-data are interwoven in order to provide support in expression generation to the domain expert.

Using Generic Meta-Data-Models for Clustering Medical Data

We present a generic, meta-model based data storage system for research, clinical studies or disease registers, which is enabled to store data of almost arbitrary structure. The system is highly costumizeable and allows the user to set up a professional web-based data acquisition system including administration area, data input forms, overview tables and statistics within hours. Furthermore, we evaluated a number of clustering algorithms regarding their ability to cluster the stored datasets for similarity search and further statistical analysis.

An ontology-based clinical data warehouse for scientific research

Medical research but also quality management is based upon medical data. The integration, validation, processing, and exploration of this data is known to be a technical obstacle for researching medical domain experts and a major pitfall to (bio‐)medical research projects. To overcome this pitfall and actively support the medical domain expert in these tasks, we present an ontology‐based clinical data warehouse for scientific research. It is completely generic and adapts itself at run‐time to the current domain‐ontology, which can be freely defined by the domain expert and describes the actual field of research. The whole system adapts is appearance and behavior to this central ontology and appears to the user like a custom made solution. Furthermore, the elaborate structural meta‐information from the ontology is used to actively support the user in tasks that usually require profound IT knowledge, such as defining complex search queries or data quality constraints, or applying advanced data visualization algorithms to the data. The proposed warehouse supports the domain expert trough the whole process of knowledge discovery from data integration to exploration.

Meta-model based knowledge discovery

Data acquisition and data mining are often seen as two independent processes in research. We introduce a meta-information based, highly generic data acquisition system which is able to store data of almost arbitrary structure. Based on the meta-information we plan to apply data mining algorithms for knowledge retrieval. Furthermore, the results from the data mining algorithms will be used to apply plausibility checks for the subsequent data acquisition, in order to maintain the quality of the collected data. So, the gap between data acquisition and data mining shall be decreased.