After a couple of changes, my colleague Lourdes Araújo and me will be presenting
multikulti algorithm in the Nova movie theater in Trondheim; it's quite cool to hold a conference in a cineplex, if only cineplexes had more wall sockets to plug the laptops in. Even if it's not in the schedule yet, our presentation will take place in
session S5-1. By now we'll have lost half the potential audience (all two of them), but we'll try to do our best anyways.
Meanwhile, I'm enjoying the conference more than I expected to; maybe it's the environment, more than the fact itself. You can follow my onsite postings (and a couple of others, at least) in
twitter
As
our mate Carlos tells in our research team blog, we have uploaded
our paper Using Dissortative Mating Genetic Algorithms to Track the Extrema of Dynamic Deceptive Functions, which will eventually become a part of his PhD thesis.
ADMGA (adaptive-threshold dissortative mating GAs) is a nifty idea: try and preserve diversity by making individuals in the population only mate with those that are
different enough. Diversity is always important in GAs, but even more so in problems where
memory is a bonus, like dynamic optimization problems.
Results obtained have been quite good, even more so with the hardest instances. So, good luck with your PhD, Carlos.
Etiquetas: PhDs, diversity, evolutionary algorithms, DOP
Time, on
withdrawal of the Spanish troops from Kosovo:
After putting so much effort into proving it belonged at the adults' table of foreign affairs, this embarrassing episode makes Spain look a little amateurish.
It's not a big deal, anyways. Long-term relationship between Spain and its natural allies will probably not be harmed. However, it does not help, either.
Etiquetas: zp, foreign policy
Our last paper on modelling time streams such as the streams of comments to this site,
Blogalia, which I have coauthored with
Dr. Lourdes Araújo,is
available online. From the abstract:
This paper presents an evolutionary algorithm for modeling the arrival dates in time-stamped data sequences such as newscasts, e-mails, IRC conversations, scientific journal articles or weblog postings. These models are applied to the detection of buzz (i.e. terms that occur with a higher-than-normal frequency) in them, which has attracted a lot of interest in the online world with the increasing number of periodic content producers. That is why in this paper we have used this kind of online sequences to test our system, though it is also valid for other types of event sequences. The algorithm assigns frequencies (number of events per time unit) to time intervals so that it produces an optimal fit to the data. The optimization procedure is a trade off between accurately fitting the data and avoiding too many frequency changes, thus overcoming the noise inherent in these sequences. This process has been traditionally performed using dynamic programming algorithms, which are limited by memory and efficiency requirements. This limitation can be a problem when dealing with long sequences, and suggests the application of alternative search methods with some degree of uncertainty to achieve tractability, such as the evolutionary algorithm proposed in this paper. This algorithm is able to reach the same solution quality as those classical dynamic programming algorithms, but in a shorter time. We also test different cost functions and propose a new one that yields better fits than the one originally proposed by Kleinberg on real-world data. Finally, several distributions of states for the finite state automata are tested, with the result that an uniform distribution produces much better fits than the geometric distribution also proposed by Kleinberg. We also present a variant of the evolutionary algorithm, which achieves a fast fit of a sequence extended with new data, by taking advantage of the fit obtained for the original subsequence.
If Springerlink is not available in your institution, please email me for a copy. It's been really a long time from the initial version, with 3 revisions (and even more if you include the versions that were sent, and rejected, from other journals), but finally it's been published. Now, on to the next one...
Etiquetas: buzz detection, evolutionary algorithms, soft computing, stream modelling
... is as relevant as ever, so our paper on Finding relevant variables in a financial distress prediction problem using genetic programming and self-organizing maps, a chapter in the book Natural Computing in Computational Finance is quite timely. This paper uses two different techniques to find out which of the dozen or so variables used in financial prediction is more relevant when predicting distress: Genetic Programming and Self-Organizing Maps (via planes analysis). It so happens there's not much overlap between what they found, but they both found a small set that synthesize financial situation quite nicely.