Report on Multi-agent Data Fusion System: Design and implementation issues 1 By Ganesh Godavari.

Slides:



Advertisements
Similar presentations
ID-04 Building Communities and Increasing Awareness & ID-06 (proposed) Building a user-driven GEOSS.
Advertisements

Chapter 11 user support. Issues –different types of support at different times –implementation and presentation both important –all need careful design.
C6 Databases.
Report on Common Intrusion Detection Framework By Ganesh Godavari.
Visual Data Mining: Concepts, Frameworks and Algorithm Development Student: Fasheng Qiu Instructor: Dr. Yingshu Li.
Robot Sensor Networks. Introduction For the current sensor network the topography and stability of the environment is uncertain and of course time is.
Mining databases with different schema: Integrating incompatible classifiers Andreas L Prodromidis Salvatore Stolfo Dept of Computer Science Columbia University.
Student : Wilson Hidalgo Ramirez Supervisor: Udaya Tupakula Filtering Techniques for Counteracting DDoS Attacks.
On the Relationship between Visual Attributes and Convolutional Networks Paper ID - 52.
Report on Intrusion Detection and Data Fusion By Ganesh Godavari.
Help and Documentation zUser support issues ydifferent types of support at different times yimplementation and presentation both important yall need careful.
1 A Framework for Measurement Valérie Paulus, Miguel Lopez, Gregory Seront, Simon Alexandre.
Ensemble Learning: An Introduction
16722 Sensing and Sensors Mel Siegel )
Distributed data fusion in peer-to-peer environment Sergiy Nazarko, InBCT 3.2, Agora center, University of Jyväskylä.
1 Enviromatics Decision support systems Decision support systems Вонр. проф. д-р Александар Маркоски Технички факултет – Битола 2008 год.
Configuration Management
Data Mining: A Closer Look
LÊ QU Ố C HUY ID: QLU OUTLINE  What is data mining ?  Major issues in data mining 2.
Overview of Distributed Data Mining Xiaoling Wang March 11, 2003.
Attention Deficit Hyperactivity Disorder (ADHD) Student Classification Using Genetic Algorithm and Artificial Neural Network S. Yenaeng 1, S. Saelee 2.
Kansas State University Department of Computing and Information Sciences CIS 830: Advanced Topics in Artificial Intelligence From Data Mining To Knowledge.
2nd Semantic Web Mining Workshop at ECML/PKDD-2002, August 2002, Helsinki, Finland Data Fusion and Semantic Web: Meta-Models of Distributed Data and Decision.
Issues with Data Mining
Comparing the Parallel Automatic Composition of Inductive Applications with Stacking Methods Hidenao Abe & Takahira Yamaguchi Shizuoka University, JAPAN.
 To explain the importance of software configuration management (CM)  To describe key CM activities namely CM planning, change management, version management.
Configuration Management (CM)
“Solving Data Inconsistencies and Data Integration with a Data Quality Manager” Presented by Maria del Pilar Angeles, Lachlan M.MacKinnon School of Mathematical.
LOGO Ensemble Learning Lecturer: Dr. Bo Yuan
Report on Intrusion Detection and Data Fusion By Ganesh Godavari.
Havva Alizadeh Ferdowsi University of Mashhad, WTLab Spring
INTERACTIVE ANALYSIS OF COMPUTER CRIMES PRESENTED FOR CS-689 ON 10/12/2000 BY NAGAKALYANA ESKALA.
Today Ensemble Methods. Recap of the course. Classifier Fusion
Exploiting Context Analysis for Combining Multiple Entity Resolution Systems -Ramu Bandaru Zhaoqi Chen Dmitri V.kalashnikov Sharad Mehrotra.
A Brief Review of Theory for Information Fusion in Sensor Networks Xiaoling Wang February 19, 2004.
Ensemble Learning Spring 2009 Ben-Gurion University of the Negev.
Institut für Softwarewissenschaft - Universität WienP.Brezany 1 Meta-Learning in Distributed Datamining Systems Peter Brezany Institut für Softwarewissenschaft.
BAGGING ALGORITHM, ONLINE BOOSTING AND VISION Se – Hoon Park.
Advanced Controls and Sensors David G. Hansen. Advanced Controls and Sensors Planning Process.
Supporting Researchers and Institutions in Exploiting Administrative Databases for Statistical Purposes: Istat’s Strategy G. D’Angiolini, P. De Salvo,
Online Construction of Analytical Prediction Models for Physical Environments: Application to Traffic Scene Modeling Anurag Umbarkar, Shreyas K Rajagopal.
Chap#11 What is User Support?
Benefits of integrating meta data into a context model Nicola Hönle, Uwe-Philipp Käppeler, Daniela Nicklas, Thomas Schwarz, Matthias Grossmann Nexus Center.
Learning to Share Meaning in a Multi-Agent System (Part I) Ganesh Padmanabhan.
1 Earth Science Technology Office The Earth Science (ES) Vision: An intelligent Web of Sensors IGARSS 2002 Paper 02_06_08:20 Eduardo Torres-Martinez –
Software Testing and Quality Assurance 1. What is the objectives of Software Testing?
Random Forests Ujjwol Subedi. Introduction What is Random Tree? ◦ Is a tree constructed randomly from a set of possible trees having K random features.
1 January 24, 2016Data Mining: Concepts and Techniques 1 Data Mining: Concepts and Techniques — Chapter 7 — Classification Ensemble Learning.
Unclassified//For Official Use Only 1 RAPID: Representation and Analysis of Probabilistic Intelligence Data Carnegie Mellon University PI : Prof. Jaime.
Emerging and Evolving Cyber Threats Require Sophisticated Response and Protection Capabilities  Advanced Algorithms  Cyber Attack Detection and Machine.
NISSC-CS Intelligence/Information Fusion Kickoff/Coordination Meeting 1/23/2004.
Pattern Recognition. What is Pattern Recognition? Pattern recognition is a sub-topic of machine learning. PR is the science that concerns the description.
Introduction to Machine Learning, its potential usage in network area,
MAIN PROJECT IMAGE FUSION USING MATLAB
The Components of Information Systems
3 Types of Data Fusion in OWS-8
An assessment framework for Intrusion Prevention System (IPS)
Information exchanges between router agents
Cost-Sensitive Learning
The Components of Information Systems
Software Measurement Process ISO/IEC
Data Mining Practical Machine Learning Tools and Techniques
Cost-Sensitive Learning
CVE.
Chapter 11 user support.
Pre-classification and AI
Information System Building Blocks
Formalization of Trust, Fraud, and Vulnerability Analysis
Modeling IDS using hybrid intelligent systems
Presentation transcript:

Report on Multi-agent Data Fusion System: Design and implementation issues 1 By Ganesh Godavari

Data Fusion Data Fusion : task of data processing aiming at making decisions on the basis of distributed data sources specifying an object Data sources –Different physical nature Electromagnetic signals, sensor data… –Different accuracy Reliability?

JDL views Data and Information Fusion –Multi level process –Level0 Fusion of sensor signals to produce semantically understandable data –Level1 Make decisions with regard to classes of objects –Level2 Asses a situation constituted by the set of aboce objects –Level3 Impact assessment i.e. adversary intent assesment on the basis of situation development prediction –Level4 Calculation a feedback like planning resource usage, sensor management etc –Level5 Human activity and situation management

Applications of DF Some applications of data fusion –Detection of intrusions into computer networks Large data available through tools like tcpdump, IDS… –Analysis and prognosis of natural and man- made disaster development Prediction and prevention of calamities like earthquakes, floods, weather conditions, nuclear explosions effect

Focus/strategy of the paper Focus –Design and implementation of DF system at Level1 Proposed strategy –Multilevel hierarchy of classifiers –Source based classifiers Decision based on data of particular sources followed by meta-level decisions

Advantages of the strategy Advantages –Decrease of the data sources information exchange –Simplicity of data source classifiers fusion even if they use different representation structures, certainty, accuracy etc.. –Use of mathematically sound mechanism for combining decisions of multiple classifiers

Problems inherent to DF applications Cause of concern –Data sources are physically distributed Spatially distributed, represented in different databases, located on different hosts –Hetrogeneous Diversity of possible data structures, difference in data structures/data specification language

Problem list Problem –development of the shared thesaurus providing for monosemantic understanding of the terminology –Entity identification problem Data specifying an object is represented in different data sources Non coherency of data measurement scales –data specifying in different sources the same entity attribute can be of different structures.

Technical terms Decision –In DF tasks its classification of an entity (object, state of an object, situation etc) Base-level/base classifiers –scheme of data fusion, each local data source is associated with a single or several classifiers

Classification of multi-level classifiers Approaches for combining decisions of multilevel classifiers can be grouped into four groups: –Voting algorithms; –Probability-based or fuzzy algorithms; –Meta-learning algorithms based on stacked generalization idea; –Meta-learning algorithms based on classifiers' competence evaluation.

Meta classification scheme

Competence based approach to combine decisions of multiple classifiers

Meta model of training and testing data Important peculiarities from learning viewpoint –Data are distributed in space and stored in different databases; –Each data source only partially specifies the same object to be classified in terms of attributes which can be different in different data sources; –Data can be incomplete; it can contain particular attribute values and also the total records in a source missed

Questions ?

References Multi-agent Data Fusion Systems: Design and Implementation Issues by Vladimir Gorodetski, Oleg Karsayev and Vladimir.Samoilov