I have a PhD in computational linguistics
and work as an independent consultant.

You can reach me at:
eric[dot]van[dot]horenbeeck[at]telenet[dot]be
My Linkedin profile: http://be.linkedin.com/in/ericvanhorenbeeck
Another subject of interest is the research of bounded growth models. Although many systems may grow exponentially for a time, no real life system is permanently unbounded. A widely used modification is the logistic form where a negative feedback term is added. The growth rate begins exponentially but decreases to zero as the population approaches the limit, producing an S-shaped trajectory. In the standard approach the parameters for the equation are deductible from the data, but the asymptote or limit is exogenous. A novel procedure to estimate also the asymptote from the data is proposed. The method consists of iteratively sorting the data in two sets: one inside and one outside the Gaussian distribution. The proposition is evaluated with consumer sales figures subjected to different promotional stimuli. In a second article it is demonstrated how buy and sell signals in a stock market can be derived based on the same concept.
A demo is available here.
My PhD dissertation explores a novel way to uncover content from an assembly of documents in an unsupervised way. The documents are added continuously (in real time) into an unrestricted text network, forcing the object description to adapt to the new facts. A text network is conceptually interesting and computationally efficient for this work. Adding data does not force the system to recalculate the entire structure. The solution proposed here constructs an intermediate layer of topical facets between the actual language production and the vocabulary. Topical facets are sets of informative phrases shared by two or more texts. Topical facets endorse access to the documents in different ways. The search for meaningful topics is one of them, auto-categorization another. A software application engaging the topical facets, demonstrates how a user is assisted to explore and to query a body of possibly unknown data.
Finding the Optimal Trade Deal and Negotiation Target
Detecting State Transitions in a Stock Market