System G Healthcare Solution
Many data mining applications such as healthcare analytics need intuitive and interactive visualization to quickly convey information and to collect feedback from users. One of biggest challenges here is how to visualize and interact with clusters of patients. We developed a suite of visual analytics methods for visualizing and manipulating clusters. Visualization and data mining are integrated together. Initially, data mining model guides the visualization to a set of relevant patients. Then, through interactions and visualization, user feedback are collected and used to update the underlying data mining models. For example, when applying patient similarity analytics, we can retrieve a set of similar patients for an index patient; using DICON visualization, users can interactively cluster those similar patients; after defining patient clusters, users can easily tell the system which clusters are relevant and which are not to the index patient; finally, our patient similar model can be updated based on the feedback.
FacetAlas visualizes heterogeneous types of patient relations
Documents in rich text corpora usually contain multiple facets of information. For example, an article about a specific disease often consists of different facets such as symptom, treatment, cause, diagnosis, prognosis, and prevention. Thus, documents may have different relations based on different facets. Powerful search tools have been developed to help users locate lists of individual documents that are most related to specific keywords. However, there is a lack of effective analysis tools that reveal the multifaceted relations of documents within or cross the document clusters. In this paper, we present FacetAtlas, a multifaceted visualization technique for visually analyzing rich text corpora. FacetAtlas combines search technology with advanced visual analytical tools to convey both global and local patterns simultaneously. We describe several unique aspects of FacetAtlas, including (1) node cliques and multifaceted edges, (2) an optimized density map, and (3) automated opacity pattern enhancement for highlighting visual patterns, (4) interactive context switch between facets. In addition, we demonstrate the power of FacetAtlas through a case study that targets patient education in the health care domain. Our evaluation shows the benefits of this work, especially in support of complex multifaceted data analysis.
SolarMap explains patient clusters by connecting clusters with the concepts on the outer rings
Documents in rich text corpora often contain multiple facets of information. For example, an article from a medical document collection might consist of multifaceted information about symptoms, treatments, causes, diagnoses, prognoses, and preventions. Thus, documents in the collection may have different relations across each of these various facets. Topic analysis and exploration for such multi-relational corpora is a challenging visual analytic task. This paper presents Solar Map, a multifaceted visual analytic technique for visually exploring topics in multi-relational data. Solar Map simultaneously visualizes the topic distribution of the underlying entities from one facet together with keyword distributions that convey the semantic definition of each cluster along a secondary facet. Solar Map combines several visual techniques including 1) topic contour clusters and interactive multifaceted keyword topic rings, 2) a global layout optimization algorithm that aligns each topic cluster with its corresponding keywords, and 3) an optimal temporal network segmentation and layout method that renders temporal evolution of clusters. Finally, the paper concludes with two case studies and quantitative user evaluation which show the power of the Solar Map technique.
ICE: Visual Analytic System for Interactive Clustering and Exploration on High-dimensional Data
Data mining operations such as similarity-based retrieval and clustering are often imperfect especially for high-dimensional data. These operations often require user interactions and feedback to adjust the results. In this paper, we present an Interactive Clustering and Exploration (ICE) system that 1) enables users to visually cluster and explore high-dimensional data, and 2) learns a distance metric based on user feedback in order to improve quality of future queries. The core algorithm behind the system is a fast interactive distance metric learning method that leverages user interactions to improve both the distance metric and dimension ranking. Building on the core algorithm, ICE system provides the following components: (1) a retrieval engine that can support similarity queries and returns clusters of similar entities, (2) an interactive visualization that presents the retrieval results and enables intuitive user interactions for cluster refinement and dimension selection, and (3) a distance metric engine that exploits user feedback to improve the underlying distance measure for subsequent queries. In the evaluation, we present the learned distance metric in ICE significantly outperforms two baseline methods. We also demonstrate the ICE system on two real use cases for music recommendation and disease risk identification.