Ibm i2
2 likes2,572 views
This document provides an agenda for a presentation on IBM's i2 solutions for fraud detection and analysis. The agenda includes introductions and overviews of IBM's solution architecture, search functions, statistics functions, geo-referenced intelligence functions, analysis functions, and time for questions. The document also provides background on IBM's portfolio of intelligence analysis products, which includes visual analysis and collaboration tools, workflow and case management tools, and analytics layers to analyze both structured and unstructured information from various sources.
Ibm i2
- 1. Industry Solutions Soluções IBM / i2 Para Detecção e Análise de Fraude © 2012 IBM Corporation
- 2. Agenda 14:00 – Introdução 14:15 – Arquitetura da Solução IBM 14:30 – Função de Pesquisa 15:00 – Função de Estatística 15:30 – Função de Inteligência Geo-referenciada 16:00 – Função de Análise 16:30 – Perguntas/considerações finais
- 3. A IBM está construindo soluções em diversos segmentos baseadas em um portifolio amplo de ferramentas More than $14B in Acquisitions 2011 Since 2005 Intelligent Investigations Financial Risk Management More than 10,000 Technical Social Analytics/Consumer Insight Professionals Workload Optimized Systems More than 7,500 Dedicated Advanced Case Management Consultants Content Analytics Largest Math Department Decision Management in Private Industry Stream Computing Pervasive Content More than 27,000 Business pureScale Partner Certifications pureXML Deep Compression Developer Productivity Autonomic Operations 2005 3
- 4. Intelligence Analysis Product Portfolio Um conjunto completo de recursos é necessário para resolver o desafio Visual Analysis & Collaboration Operational dashboards, workflow, case adjudication Workflow & Case Management Visualization & Operational Link Analysis Dashboards Analytics Layer Intelligence, Descriptive & Predictive Analytics against structured, semi-structured and unstructured information Descriptive and Content & Streams Predictive Sentiment Analytics Analytics Analytics Information Sharing Trusted Information Layer Analysis Repository Establish, Manage, Share & Deliver information that is accurate, complete, in context and insightful Persistent Relationship Awareness ETL Data on Demand Data Quality Data Delivery Source Systems Structured & Unstructured Content
- 6. Watson Today: Processes Unstructured Text & 200 Hypotheses/3 seconds Hypothesis Generated from “Answer” Answer: A large country in the Western Hemisphere Static Guess Questions Q1, Q2 … Qi whose capital has a similar name. Data Corpus Statistical Ensemble S1 S2 S3 ... SN of 600 to 800 Scoring Engines Watson 3,000 cores;100 TFlops 2 TB memory ~ 200 KW Element Refresh Time ~30 Machine Learning Models Weigh Data Corpus 2 Weeks Evidence-Based Scores, Produce Confidence for Each Decision Question 0<P<1 Hypothesis Engines Weeks to Months Support System Hypothetical Question With Scoring Engines Weeks to Months Greatest Confidence is Chosen Decision 4 Days Support Engine Question: What is Brazil? 6
- 7. Questions?