Data Mining:
What is it?
Who uses it?
In engineering &
  manufacturing

In financial services

Enterprise Risk Management

Complexica Products and Services

The types of products/services we can provide come in several forms:

Consulting

We can do your advanced data mining for you. We can provide data mining services for the client. This includes the development of "credit scores" or "marketing scores" that can be implemented in the client's credit approval, portfolio management, or marketing campaign process. The scores can be installed by the client or by us.

Data Mining Software

We build software that can be installed on the client site. This includes software that can be used by the client's analytic group. The most up-to-date as well as time-tested techniques can be included. The software is customized to the client's particular analytic needs. The form the software takes depends on the client needs. For small data flow problems, the interface is written in Excel while the analytic engines are written in C++ or other statistics packages that interface with Excel, such as S-Plus.

As required we can add the following analytic modules:

  1. linear regression -- a standard statistical technique for curve fitting and trend analysis.
  2. logistic regression -- a standard statistical technique for classification. This technique is often used to identify which are good and which are bad customers.
  3. Partial Least Squares -- a regression technique that can be used to reduce the amount of data a client needs to collect.
  4. Sliced Inverse Regression -- a simple nonlinear method for curve fitting and trend analysis.
  5. Fisher Discriminants -- a method for segmentation of data, often used to identify customer segments.
  6. Neural networks -- a technique for extracting patterns from data that is based on the way human brains make sense of their environments.
  7. Radial basis functions -- a technique that looks for patterns in similar data.
  8. Genetic Algorithms -- an optimization method based on natural selection and evolution.
  9. Simulated Annealing -- an optimization method that is based on the physics of cooling metals.
  10. Advanced clustering -- for customer segmentation.
  11. CART and other decision trees -- for customer segmentation.

The documentation for the software is online and the software itself is user friendly and designed in many cases to sit on the decision-maker's desk.

Training

We build a data mining capability within the client. We provide the necessary training to take full advantage of both the consulting services and the analytic software. This training can be one-on-one, classroom, or train-the-trainer type training. The training is customized to the client's needs. Training sessions can be on-site or off-site in Santa Fe. Sessions can include up to 10 people and, depending on the technical sophistication of the students, can last one to five days. Often, software can be installed on the student's laptop computers.

Support

We provide ongoing software support, upgrades, and training. We answer questions from the users, identify and fix bugs, and provide the client with the latest upgrades by email when possible and by phone when necessary. We provide a 90-day guarantee. A maintenance contract is available after that.