Download Data Mining and Predictive Analytics in Various Industries and more Quizzes Information Technology in PDF only on Docsity!
TERM 1
Cabela's
DEFINITION 1 Goal- Wanted a single viewof customers across multiple channels to better focus marketing efforts to increase sales. Problem- stored data into data marts that took 1-2 weeks to just build the data Solution- Teradata and SAS integration (best-of-breed) created tools and techniques aimed at improving speed and accuracy for predictive and explanatory models Results- helped improve return on marketing, select optimal site locations, understand value of customers, design efficient promotions, Tailor direct marketing offers to customers TERM 2
Infinity P&C: Smarter
Insurance
DEFINITION 2 Problem-One out of 5 claims is fraud. Used predictive analytics to fast-track four and close cases quickly Challenge Better customer relationship management, detect fraud, & higher Return on investment Solution used predictive analytics. Rules for rating and identifying potential frauds (red flag claims) Results Improved customer service, combat fraud, and increase profit, more efficient workflow, and improved customer retention. Used IBM SPSS for their analytics tools TERM 3
Memphis PD: Analytics to combat crime
DEFINITION 3 Challenges- Rising crime, impatient city leaders, budget pressures Solution- Operation Blue CRUSH project (Crime Reduction Utilizing Statistical History). With IBM SPSS Modeler at its heart that enables officers to unlock the intelligence hidden in the departments huge library of crime records and police reports going back nearly a decade. Result- Reduced crime TERM 4
Mine on terrorist funding
DEFINITION 4 How can data mining be used to fight terrorism? Track funding of terrorists activities via Normal financial transactions Price deviations Money laundering; overvaluing imports & undervaluing exports Tax avoidance or evasion Does it jeopardize individual privacy? Yes, perhaps by tracking personal and financial data of individuals TERM 5
Data Mining goes to Hollywood
DEFINITION 5 Problem- predicting success of movie is challenging and complex due to the difficulty with forecasting demand. Proposed solution- Researcher Ramesh Sharda and Dursun Delen used a classification model into one of nine categories. Used methods like neural networks, decision trees, support vector machines and three types of ensembles. Used IBM SPSS Modeler. Results- Used the data from all three ensembles (SVM, ANN, CART) which lead to a significantly low standard deviation obtained from the ensembles compared to the single models.
TERM 6
Target: predicting Customer buying patterns
DEFINITION 6 Target sent a teen maternity ads because Target's analytic model suggested she was pregnant based on her buying habits. Target found out teen girl was pregnant before her father knew! Tradeoff between knowledge discovery and privacy rights. Target did not violate her rights, they just did what most retailers are doing now and tracking transactional data to predict information about their customers. Risk offending customers and hurt the bottom line. TERM 7
Watson: Man versus Machine
DEFINITION 7 Situation- Can a machine beat the best of man in what man is supposed to be best at? Problem- to build a computer system that could compete at the human champion level of Jeapardy. Solution- IBM System called DeepQA which is a massively parallel, text mining-focused, probabilistic evidence-based computational architecture. (NOT connected to internet) Results-Watson is performing at human expert levels in terms of precision, confidence, and speed at the Jeapardy quiz show TERM 8
Patent Analysis
DEFINITION 8 patent= "Exclusive rights granted by a country to an inventor for a limited period of time in exchange for a disclosure of an invention" How to do PA? via use of analytical techniques to extract valuable knowledge from patent databases Why do it? Enable competitive advantage, critical BI decisions, identifying new talent, find unauthorized use of patent, prevent competitors from creating similar products Uses TM tools to continuously dig deep into various data resources to develop a holistic view of the competitive environment TERM 9
Mining for Lies
DEFINITION 9 Problem- humans tend to perform poorly at deception- detection tasks. Humans are only 54% percent accurate in making veracity(habitual truthfulness) determinations. Study analyzed statements from people involved in crimes on military bases Message feature mining- text based deception detection method that analyzed testimonies for POI. used features(cues) TERM 10
Premier Bank Card
DEFINITION 10 Problem- old system couldn't keep up, consisted of Access and excel, manual analysis, long time, high error, low accur. Solution- Microsoft Business Intelligence(end-to-end). Users now have instant response and can pull data down out of cubes and DW. BI Stackconsists of SQL server 2005, Microsoft Analysis Services from SQL server, DW with 12 terabytes of data, SharePoint, PerformancePoint, and Office 2007 Flexible and self-service, more accurate info, saves time, saves money due to higher accuracy.
