Internet of Things, data management for healthcare applications. Ontology and automatic classifications

Similar documents
Why are Organizations Interested?

Building a Data Quality Scorecard for Operational Data Governance

Populating a Data Quality Scorecard with Relevant Metrics WHITE PAPER

STAR WARS AND THE ART OF DATA SCIENCE

A Systemic Artificial Intelligence (AI) Approach to Difficult Text Analytics Tasks

Delivering Smart Answers!

Text Analytics Evaluation Case Study - Amdocs

Big Data Text Mining and Visualization. Anton Heijs

Search and Information Retrieval

Auto-Classification for Document Archiving and Records Declaration

Measure Your Data and Achieve Information Governance Excellence

Text Analytics Software Choosing the Right Fit

What do Big Data & HAVEn mean? Robert Lejnert HP Autonomy

Hexaware E-book on Predictive Analytics

Maintaining a Competitive Edge with Interaction Analysis

Taxonomies for Auto-Tagging Unstructured Content. Heather Hedden Hedden Information Management Text Analytics World, Boston, MA October 1, 2013

White Paper. How Streaming Data Analytics Enables Real-Time Decisions

HiTech. White Paper. A Next Generation Search System for Today's Digital Enterprises

Social Media Implementations

IDC MaturityScape Benchmark: Big Data and Analytics in Government. Adelaide O Brien Research Director IDC Government Insights June 20, 2014

Data Mining with SAS. Mathias Lanner Copyright 2010 SAS Institute Inc. All rights reserved.

Find the signal in the noise

EMC DOCUMENTUM CONTENT ENABLED EMR Enhance the value of your EMR investment by accessing the complete patient record.

Master of Science in Health Information Technology Degree Curriculum

Data Mining Analytics for Business Intelligence and Decision Support

Data Sheet: Archiving Symantec Enterprise Vault Discovery Accelerator Accelerate e-discovery and simplify review

SPATIAL DATA CLASSIFICATION AND DATA MINING

ENTERPRISE DOCUMENTS & RECORD MANAGEMENT

BUSINESS VALUE OF SEMANTIC TECHNOLOGY

UTILIZING COMPOUND TERM PROCESSING TO ADDRESS RECORDS MANAGEMENT CHALLENGES

CONCEPTCLASSIFIER FOR SHAREPOINT

Promises and Pitfalls of Big-Data-Predictive Analytics: Best Practices and Trends

STATISTICA. Clustering Techniques. Case Study: Defining Clusters of Shopping Center Patrons. and

Text Mining and Analysis

TEXT ANALYTICS INTEGRATION

The University of Jordan

Auto-Classification in SharePoint. How BA Insight AutoClassifier Integrates with the SharePoint Managed Metadata Service

IBM Content Analytics with Enterprise Search, Version 3.0

Governance in Digital Asset Management

Data Governance. David Loshin Knowledge Integrity, inc. (301)

The Business Value of Predictive Analytics

Flattening Enterprise Knowledge

Getting Started with Data Governance

EHR Standards and Semantic Interoperability

User Needs and Requirements Analysis for Big Data Healthcare Applications

Semantic Data Management. Xavier Lopez, Ph.D., Director, Spatial & Semantic Technologies

EUCISE2020 Industry Day

Standardization Requirements Analysis on Big Data in Public Sector based on Potential Business Models

Transformation of Free-text Electronic Health Records for Efficient Information Retrieval and Support of Knowledge Discovery

IDC MaturityScape Benchmark: Big Data and Analytics in Government

Collaboration. Michael McCabe Information Architect black and white solutions for a grey world

Improving the Performance of Data Mining Models with Data Preparation Using SAS Enterprise Miner Ricardo Galante, SAS Institute Brasil, São Paulo, SP

US Department of Education Federal Student Aid Integration Leadership Support Contractor January 25, 2007

Microsoft FAST Search Server 2010 for SharePoint Evaluation Guide

Enterprise Data Quality Dashboards and Alerts: Holistic Data Quality

KPMG Unlocks Hidden Value in Client Information with Smartlogic Semaphore

Big Data & Security. Aljosa Pasic 12/02/2015

Industry Models and Information Server

The Value of Taxonomy Management Research Results

2011 Cyber Security and the Advanced Persistent Threat A Holistic View

Developing Microsoft SharePoint Server 2013 Advanced Solutions

Facilitating Business Process Discovery using Analysis

Healthcare Measurement Analysis Using Data mining Techniques

System Behavior Analysis by Machine Learning

An Introduction to Data Mining

Guideline for Implementing the Universal Data Element Framework (UDEF)

Apigee Insights Increase marketing effectiveness and customer satisfaction with API-driven adaptive apps

How To Understand The Difference Between Terminology And Ontology

Get More Value from Your Reference Data Make it Meaningful with TopBraid RDM

A Capability Model for Business Analytics: Part 2 Assessing Analytic Capabilities

A Survey on Web Mining From Web Server Log

CAPTURING THE VALUE OF UNSTRUCTURED DATA: INTRODUCTION TO TEXT MINING

Developing Microsoft SharePoint Server 2013 Advanced Solutions MOC 20489

Hurwitz ValuePoint: Predixion

B.Sc. in Computer Information Systems Study Plan

This software agent helps industry professionals review compliance case investigations, find resolutions, and improve decision making.

ECM Governance Policies

Extend your analytic capabilities with SAP Predictive Analysis

Questionnaire on the European Data-Driven Economy

WHITE PAPER. Creating your Intranet Checklist

How To Perform An Ensemble Analysis

IBM SPSS Modeler Premium

Certified Information Professional 2016 Update Outline

Health Data Analysis Specialty Track Curriculum Competencies

Building Data Cubes and Mining Them. Jelena Jovanovic

Survey Results: Requirements and Use Cases for Linguistic Linked Data

OLAP and Data Mining. Data Warehousing and End-User Access Tools. Introducing OLAP. Introducing OLAP

ANALYTICS IN BIG DATA ERA

NICE MULTI-CHANNEL INTERACTION ANALYTICS

IMPROVEMENT THE PRACTITIONER'S GUIDE TO DATA QUALITY DAVID LOSHIN

Voice. listen, understand and respond. enherent. wish, choice, or opinion. openly or formally expressed. May Merriam Webster.

Transcription:

Internet of Things, data management for healthcare applications. Ontology and automatic classifications Inge.Krogstad@nor.sas.com SAS Institute Norway

Different challenges same opportunities! Data capture in value chain Information available across value chains Share same understanding of data and content Globalization and universal access to information Empower analytics for right time decision making

Do we find the IoT in the strategic plans for Norwegian Healthcare?

Main objectives in the strategic plans for Norwegian Healthcare! Improve quality and interactions joint goals in healthcare and care sector Overall System of concepts to support continuity of care Improve healthcare and care sector through information technology

Improvement areas in the strategic plans for Norwegian Healthcare! Diffuse management model The organization is not adapt for interaction Large variation on productivity Private healthcare is more effective Weak quality measures Increasing proportion of unstructured data SOURCE:

How can IoT potentially fit into the strategic plans? Value / Cost efficiency RFID areas Logistics for patients and staff Logistics for equipment and supplies Security systems Tracing and tracing objects Tracing and tracing patients Maintenance and implementation Improve quality and interactions joint goals in healthcare and care sector Overall System of concepts to support continuity of care Improve healthcare and care sector through information technology Diffuse management model The organization is not adapt for interaction Large variation on productivity Private healthcare is more effective Weak quality measures Increasing proportion of unstructured data Today 2013

Some challenges and possible data management minefields Value / Cost efficiency Right to privacy Diffuse management model The organization is not adapt for interaction Large variation on productivity Private healthcare is more effective Weak quality measures Increasing proportion of unstructured data RFID areas Logistics for patients and staff Logistics for equipment and supplies Security systems Tracing and tracing objects Tracing and tracing patients Maintenance and implementation RFID tag prize pr. unit Improve quality and interactions joint goals in healthcare and care sector Overall System of concepts to support continuity of care Improve healthcare and care sector through information technology Lack of standardization Today 2013

Deviation from plan The data management challenge - interoperability Transactions + Org. unit A Org. unit B Org. unit C Org. unit D Org. unit E Electronic Health Record systems Theoretical progress - Staff data Patient data Equipment Maintenance Same format Mutual understanding of content Supply Activities

Interoperability Operational data Knowledge and Quality Data Key dimensions of interoperability Knowledge and skills Business processes and value chain ICT, data, applications and communication Semantic, definitions and insight Value creation through interoperability Semantics Business processes ICT Knowledge SERES, The register of semantic for electronic collaboration SEMICOLON, ICT-based methods, tools and metrics for semantic and organizational interoperability

The interoperability challenge related to data management Operational data Knowledge and Quality Data Data integrations Data feeds

The interoperability challenge related to applications Data feeds Enterprise End-user Application Automated or semiautomatic data feeds Decision processes Anaesthesia system Surgery planning Diagnostic Pain therapy Manually data entry Automatic data capure Work processes Surgery Transportation Medical treatment Births Databases did not bring data into structure 80% of data in healthcare is unstructured Unstructured data is increasing 60 % per. year

Text Analytics Information Organization and Access Predictive Modeling, Discover Trends and Patterns Content Categorization Ontology Management Text Mining Sentiment Analysis

Categorization Determine topics / subject area(s) of a particular document Example Relevance Why accessing a previous patioent in the Electronic Health Record systems? Associate rules to a category Example Reversing treatment D-vitamin is indicator for wrong medical treatment for diagnosis group Statistical or Rule based definition of topics Example Professional area Only above P20 is relevant for knowledge building Rule based types: Linguistic or Boolean Example Category matches if the sum of weights of terms exceeds certain threshold

Content Categorization, Entities Extraction, Fact and Event Extraction Automatic Categorization Map documents to one or more topics according to a taxonomy Taxonomy Management Design, test and development of a set of topics (taxonomy) Design automatic categorization rules Collaboration allowing several knowledge experts to work together Entities Extraction Find entities in text: people, location, companies, Fact and Event Extraction Extraction of relations between entities

Categorization Testing Multiple document formats supported (TXT, PDF, XML, HTML, RTF, etc.) Test documents are used to verify the performance of a rule Well performing rule will match all of the relevant test documents (recall) while not matching irrelevant documents (precision) Results are PASS/FAIL Fail: Document is NOT part of this group Documents Categorizer Pass: Document is part of this group

Analysis of Unstructured Data Integration of Text Mining and Content Categorization Enterprise Content SAS Text Miner Automated Discovery of Text Structure Content with Metadata SAS Enterprise Content Categorization Expert-based Refinement of Metadata Content with optimized Metadata SAS Enterprise Content Categorization SAS Text Miner Enterprise Content Expert-based Definition of Metadata Content with Metadata Automated Discovery of Text Structure with additional Metadata Content with optimized Metadata

SAS Ontology Management Build semantic repositories to manage companywide thesauri, vocabularies, and build relationships between them Create structure for integration with structured data and Contextual Analysis Maintain metadata across repositories and databases and to automatically tag documents according to the defined taxonomies Simplify the task of obtaining and returning knowledge from input documents

SAS Ontology Management Enables collaborative ontology development and maintenance Integrates existing document repository assets Identifies relationships between document repositories Build subject-matter expertise into search-and-retrieval activities Consistently applies subject-matter expertise across document repositories in real time Centralizes administration for collaborative ontology development

Examples Classification of Electronic Health Record data Rescue team and alert planning Detecting unauthorized access to patient data

Bringing information together Search and Summarization Electronic Health Record systems (EPJ)

Example Taxonomy

A study of Location of Rescue Teams, RT optimization

i I Open i LocationsOpen A Real time study on location data and demands Problem Formulation and Solving Objective Function: Minimize the maximum distance between stations and areas subject to: Max Distance Definition: define the largest distance (1) MaxDist Distance ij X ij Staffed Constraint: total stations with ambulances must equal the number available to open (2) Σ i Open i = Number of Ambulances Service Constraint: Stations that cover an area must have an ambulance assigned to them (3) X i j Open i for all i, j Cover Constraint: sum of coverage must meet demand (4) Σ i X ij = Demand j Supply Constraint: sum of coverage must not exceed supply (5) Σ j X ij Supply i Open i

Pattern recognition Detect unauthorized access to data in Electronic Health Record systems (EPJ), Association analysis Clustering MBR (K-nearest neighbors) Link analyses Dynamic building rules based on classification, profiling and white lists Access logs and EHR are analyzed through scenarios and scoring process Data intergartion Investigation Transformation and algoritmes Detection and scoring Desicion making Analyses White lists and scenarios

Test on Wk0001

Solving data management IoT enables a potential for value creation! Ability to define a hierarchical taxonomy where related topics are grouped together (Identify enterprise and structure according to ISO/CEN) Implement automatically classifies of documents using customizable rules for precise categorization new material to existing text sources (increase with 60%/year) Establish knowledge services (SOA) that extract, discover and predict knowledge from multiple text documents (i.e. including epicrisises contents can be added to structured data) Automated or semiautomatic data feeds Cluster documents, i.e. Electronic Health Record systems (EPJ), into related groups for descriptive or predictive modeling for operational risk analysis or performance monitoring of HF s and transparency between RHF s Ability for maintain ontology in enterprise content repositories and databases. The ontology can become the key element to integrate the Clinical Decision Support system with the new National health registries Manually data entry Automatic data capture Enable new services for semantic terms that are used to organize previously disassociated and isolated text repositories i.e. data from other specialized systems (e.g. at Ullevål there are more then 200 small special systems) Establish an enterprise semantic model for Norwegian healthcare that creates and maintains consistent and centralized metadata across all structured and non-strucured data collections (ref. Samspill 2.0 and Gode helseregistre bedre helse? )

Copyright 2006, 2007, SAS Institute Inc. All rights reserved.