Workshop on Advanced Predictive Models in Healthcare
June 24th, 2017, Vienna, Austria
Health systems worldwide are under pressure to deliver better care for more people, using fewer resources. These must therefore revolutionise their ability to adapt to the needs of the populations they serve. An essential step in this transformation is the accurate and reliable prediction of health outcomes before they actually occur, both short-term and long-term, creating windows of opportunity to prevent illness and reduce care consumption through targeted interventions.
Predictive (or prognostic) models are analytical models used to predict health outcomes in defined populations. Various methodologies, developed by both the statistics and computer science communities, are used in predictive model development. To date, linear, logistic and Cox regression models are most commonly used. Recently, deep learning methods have sparked new approaches to health outcome prediction. At the same time, there are challenges posed by new data sources, such as routinely-collected clinical data in electronic health records and personal health data collected with wearable sensors.
This workshop will focus on key topics related to advanced predictive models, capable of providing actionable and timely insights. We would like to invite researchers from both academia and industry to participate in this workshop, share their opinions and experience, as well as discuss future directions.
The objectives of this workshop are:
- Bring together researchers (from both academia and industry) as well as practitioners to present their latest problems and ideas.
- Attract healthcare providers who have access to interesting sources of data and problems but lack the expertise in data mining to use the data effectively.
- Enhance interactions between data mining, text mining and visual analytics communities working on problems from medicine and healthcare.
Topics of interest
Topic areas for the workshop include (but are not limited to) the following:
- Novel approaches for prediction of health outcomes (e.g. deep learning);
- Methods for predicting from streaming data and real-time prediction methods
- Methods for handling multi-resolution longitudinal data (e.g. low-res biomarkers + hi-res wearable sensor data);
- Machine Learning methods for time-to-event prediction;
- Methods for tailoring predictive models to local populations;
- Dealing with missing data;
- Modelling and prediction of treatment response and disease trajectories;
- Methods for adjusting for treatment pollution in observational datasets;
- Methods for outcomes prediction from Electronic Health Record (EHR) data;
- Models that allow users (e.g. clinicians, patients) to explore "what-if" scenarios;
- "White box" models, i.e. helping users understand complex predictive models;
- Comparative studies of diverse prediction methods.
- Ameen Abu-Hanna, University of Amsterdam, The Netherlands
- Joydeep Ghosh, University of Texas, Austin, USA
- Tudor Groza, Garvan Institute of Medical Research, Australia
- Robert Moskovitch, Ben-Gurion University of the Negev, Israel
- Mykola Pechenizkiy, Technical University Eindhoven
- Pedro Perreira Rodrigues, University of Porto, Portugal
- Stephen Swift, Brunel University, London, UK
- Suzanne Tamang, Stanford University, USA
- Allan Tucker, Brunel University, London, UK
- Fei Wang, Cornell University
- Marinka Zitnik, Stanford University, USA
- Blaz Zupan, University of Ljubljana, Slovenia
Extended Paper Submission Deadline: 12th May 2017
- Notification of Acceptance: 26th May 2017
- Camera Ready Paper Due: 31st May 2017
- Workshop: 24th June, 2017
Papers should be submitted to the Easy Chair Website (https://easychair.org/conferences/?conf=wapmh2017). The conference features two categories of papers (please note if you are submitting a paper as a student):
- Full research papers (up to 10 pages)
- Short papers (up to 5 pages) and Demo papers (1 page) describing either:
- a short research project,
- a demonstration of implemented systems, or
- late-breaking results (work-in-progress).
Papers should be formatted according to Springer's LNCS format (see www.springeronline.com/lncs).
The papers will be included in working notes to the workshop that will be handed out to participants (electronically) . In addition, the authors of the best submissions will be invited to contribute to a special issue in the Springer’s Journal of Health Informatics Research via a fast track review.