[Talks@IDSIA] Chart Pattern Recognition using Markovian Models, talk by Yann Soullard (Friday 15 Jan, 11am)

Dear All, Yann Soullard, from the Heudiasyc Lab of the University of Compiègne will give a talk on "Chart Pattern Recognition using Markovian Models" this Friday, the 15th, at 11am in IDSIA’s meeting room (Galleria 1 bldg in Manno). Please find the abstract here below. Abstract. A chart pattern is a particular shape which appears in financial stock markets. It is defined by theoretical rules and it has a predictive power. Detecting such patterns is difficult. There is an important gap between theory and practice; real patterns do not perfectly respect the theoretical rules. Moreover, the definition of chart patterns seems subjective; it depends on the financial expert. Finally, there are no large labeled datasets of chart patterns. We investigate chart pattern recognition using statistical Markovian systems. We focused on generative (Hidden Markov Models) and discriminative (Hidden Conditional Random Fields) approaches which are standard technologies for sequential data recognition. We proposed various strategies to learn accurate systems with small labeled training sets. The first one blends HMMs and HCRFs in such a way that the modeling ability of the generative models is used to limit the overfitting of the discriminative ones. The second strategy is a semi-supervised approach which combines an HMM and an HCRF system during the learning step. This second approach has some similarity with a co-training algorithm. Best regards, Alessandro _________________________________ Alessandro Antonucci IDSIA Dalle Molle Institute for Artificial Intelligence Via Cantonale (Galleria 2) CH-6928, Manno-Lugano, CH mail: alessandro(a)idsia.ch skype: alessandro.antonucci tel: +41 916108515 web: www.idsia.ch/~alessandro _________________________________