Intellectual Properties And Other Assets

• Statistical Expertise
  • Principal statistician: Ph.D., Post-doctorate at Cambridge University, 12 years of experience in fraud detection, marketing, finance and the Internet.
  • Keywords: predictive modeling, time series, spatial processes, machine learning, decision trees, tree forests, boosted trees, regression, pattern recognition, clustering, density estimation, discriminate analysis, customer segmentation, survey analysis, large data sets, data mining, text mining, web mining, scoring technology, Monte Carlo, six sigmas, Pareto analysis, extreme values.
  • Software: SAS, Splus, R, JMP, SQL, Perl, C, C#, Java, MySQL.

• Domain Names (some of them can be rented or acquired)

  List available upon request.

• White Papers
  • Stock portfolio management for the lazy trader - currently under revision
  • Efficient risk management for hedge fund managers - PowerPoint presentation featuring original pattern recognition technique
  • Creation of optimum fraud tables for real time credit card fraud detection
  • Statistical methods for purposes of detecting click fraud - 31 pages PowerPoint presentation comparing various statistical methodologies
  • Click fraud detection - features case studies, IP blacklist, Perl script, best practices
  • Optimization of A/B testing - case study: LowerMyBills.com
  • Secret weapons to beat the big advertisers at the PPC game - provides details about our marketing and advertising weapons, currently under revision
  • Maps, zipcode data and palettes with R - explains how to produce graphs such as our crime map

• Data Sets

  • Datashaping.com Server Logs
    Contains 1.5 million hits, more than 100,000 unique IPs, traffic from several hundreds search engine distribution partners spanning over a period of more than a year. This data set features click activity from several test campaigns. Some of these campaigns (by design) contain nearly 100% click fraud while others (by design) do not contain any fraud. Part of the traffic originates from PPC advertising as we bought more than 10,000 keywords on various search engines. The data set is particularly useful to identify click behavior from bad distribution partners and when designing click fraud rules.
  • California Business Listing
    More than 700,000 businesses incorporated between 2000 and 2004, with business name, address, date of incorporation and status (LLC or INC).
  • Website Rankings Database
    Contains 1.8 million unique domains found in Google and Yahoo directories. Each domain is categorized and ranked using a proprietary algorithm.
  • IP Blacklist
    Suspicious IP addresses and IP ranges reported by webmasters since 09/01/2005 - about 500 new entries added every quarter. Also includes more than 30,000 IP addresses detected as fraudulent in various advertising campaigns. Each address has a time stamp and a score. The list is updated every month.

• Proprietary Technology (Algorithms)
  • Fraud detection: IP address scoring and IP blacklist system, logical click scoring (click fraud), unique user identification in server logs, statistical scoring system (click fraud), proprietary metrics for fraud detection.
  • Query Intelligence: keyword clustering and query correlation tables.
  • Innovative advertising technologies: click here for details. Includes large scale automated detection of under-bought and over-sold keywords on major PPC search engine programs, to further optimize PPC campaigns.

• Mailing Lists (available for rent, e.g. to publish a job ad)
  • Job e-mail alert (several thousand subscribers)
  • Statistical seminar mailing lists (individual and corporate)

• Clients / Worldwide Network
  • Click here for details. Includes Visa, Wells Fargo, InfoSpace. Clients / Partners in North America, Europe and Australia.

• Provisional Patents, Trademarks:
  • Click Scoring System to Assess Traffic Quality and Click Validity in Search Engine Networks (Provisional Patent)
  • Fast Feature Selection Algorithm to Create Robust Rules and Tree Forests for Fraud Detection, Click Scoring or Click Segmentation in Pay-Per-Click Programs (Provisional Patent)
  • Optimum Click Scoring System for Click Segmentation when Conversion Metrics are Available (Provisional Patent)
  • Contrarian Scoring Technology to Assess Internet Traffic Quality, with Application to Click Segmentation and Click Fraud Detection (Provisional Patent)

• Applications
  • Ridge regression software - source code: Perl and C version
  • Copyright and trademark infringement detection - web application
  • Query Suggestion Box - web application
  • Referral advertising technology - currently in production mode
  • Web Site Ranking Technology - web application
  • Stock market trading signals - web application

• Web Traffic
  • DataShaping.com is one of the most visited internet properties in its category, with traffic levels above or similar to Insightful.com (creator of Splus), KDNuggets.com, The American Statistical Association and top statistical job boards, with over 10,000 highly targeted unique users per month. There are currently four spots at the top of the page (across the entire website) available to advertisers.
  • Extensive PPC keyword listing and optimized PPC ads distributed across various networks.
  • Highly cost-effective proprietary mechanisms to drive traffic to datashaping.com.

• Human Resources: access to talented Consultants in US and India.

• Web Publishing Company: Language Shaping Solutions.