C&F Research

Dr. Fiorini is head of R&D at Carrera & Fiorini. At Carrera & Fiorini. We bring to your organization years of solid practical marketing experience as well a strong theoretical grounding that can expand your business in ways you never thought possible.

He has many years of experience in private industry as well as experience as a research faculty member at the University of Southern Maine from 2001-2007.

The following summarizes a portion of the research and development (R&D) at Carrera & Fiorini under the supervision of Dr. Fiorini.

Internet Marketing

• C&F is actively engaged in this area. We are always developing new mathematical algorithms that are able to quantify and analyze marketing efforts. In particular, we have been for many years researching and implementing traffic algorithms that are able quantify the impact that Search Engine Optimization, Pay-Per-Click (PPC),  (and more recently) Social Media Optimization (SMO) campaigns have on your internet marketing efforts. This enables us to more effectively develop a strategy that has a high ROI for your business.

Information Retrieval (IR)

• We are active in the area of  IR research. We have developed our own search engines and proprietary ranking algorithms.

Data Mining

• Off and on, for over 15 years, we’ve been active in the area of data mining. Sometimes it’s related to “mining” information regarding marketing or computer performance data sets (via statistical, clustering, decision trees, whatever), or developing algorithms to detect marketing behaviors and signals.

Artificial Intelligence

• We do applied research in the area, which has mostly been the application of  well understood technologies such as Neural Networks (e.g., Backpropagation, SOM, etc.), Bayesian Networks, Genetic Algorithms, Decision Trees, Reinforcement Learning, symbol-based learning algorithms to aspects of Internet marketing measurement and modeling.

Computer Performance Modeling & Evaluation

• Our research interests include modeling and evaluating the effects of power-laws on computing systems. It turns out they are everywhere! We have also done a lot of research in computer dependability and reliability theory.

• Our most interesting work in this area is solving a thought to be unsolvable problem – the RESTART problem. The difficulty here was that there was no (general) formulation characterizing the distribution of tasks, jobs, etc., that have to RESTART from the beginning again when they fail to complete. It turns out that jobs that exhibit this behavior have completion time distributions that exhibit a power-law. This has a whole bunch of applications with respect to the performance of common computer networking protocols. For recent applications of our fundamental results, see Predrag R. Jelenkovi, Columbia University.

A series of our papers that we have written related to the RESTART problem have been   recently accepted to Mathematics of Operations Research - Asymptotic Behavior of Total Times For Jobs That Must Start Over If a Failure Occurs (with Soren Asmussen, Lester Lipsky, Robert Sheahan, and Tomaz Rolski).

Linear Algebraic Queueing Theory

• Another area of research is in the area of the Linear Algebraic approach to Queueing Theory (LAQT). Typically, we utilize this when we are requested to do performance modeling.